VISUAL DETECTION OF TARGETS; ANALYSIS AND REVIEN · NAVWEPS REPORT 8617 NOTS TP 36U5 COPY 6 2...

NAVWEPS REPORT 8617 NOTS TP 36U5

COPY 6 2

VISUAL DETECTION OF TARGETS; ANALYSIS AND REVIEN

u by

Ronald A. Erickson

Aviation Ordnance Department

££-iF CO.';

L:::.-

*n

ABSTRACT. This report discusses many o^ the aspects of air-to-ground visual search for targets. Curves are presented that can be used for estimating the probability that a ground target is within view and for determining the angular rate of the target as measured with respect to the air observer. Optical aspects (clouds, atmospheric attenuation, reflectance factors) of visual detection are discussed briefly and references from which data can be obtained are cited. A number of laboratory experiments concerning visual detection are described, and some of the results are given. Examples of simulation, operational, and mathematical methods of obtaining estimates of search performance are given and compared.

U.S. NAVAL ORDNANCE TEST STATION

China Lake, California

m ̂liLuo^ lyj

February 1965

D)

.. V

l.M-U If

I^JÜ

uks ,H A

^VAL ORDNANCK TKST STATION

AN ACTIVITY OF THE BUREAU OF NAVAL WEAPONS

J. I. HARDY, CAFT., USN WM. B. MCLEAN. PH.D. Commoner Ttchnical Director

FOREWORD

This study has b««n psrforaed as part of the program of the Naval Ordnance Test Station in the field of weapon system development and analysis for conventional warfare.

This work has been supported by the Bureau of Weapons WepTask RAV32-N001/216-1/F008-02-O02 and was conducted from January 1962 to January 1963.

This report has received technical review by Roy Dale Cole, Con- sultant in the Aviation Ordnance Department; and Lewis 0. Erwin, of the Weapons Analysis Group.

Information presented herein may be subject to revision as further data become available.

Released by Under authority of N. E. WARD. Head, WM. B. NcLEAN Aviation Ordnance Department Technical Director 19 November 1964

NOTS Technical Publication 3645 NAVWEPS Report 8617

Published by Aviation Ordnance Department Manuscript , 35/MS-217 Collation Cover, 30 leaves, abstract cards First printing 1H0 numbered copies Security classification UNCLASSIFIED

ii

BLANK PAGE

NAVWEPS REPORT 8617

CONTENTS

Introduction 1 Geometry of Air-to-Ground Search 1

Aircraft Obstruction of View 1 Terrain Obstruction of View 3

Kinematics of Air-to-Ground Search * 6 Angular Rate of the Ground 7 Equivalent Angular Rates 9 Instructions for Use of Nomograph 10 Example 10 Target Angular Rate 14 Moving Target 17 Operational Effects 18

Optical Aspects of Air-to-Ground Search 19 Clouds 19 Atmospheric Attenuation 19 Terrain and Target Reflectance 20

Psychophysics of Air-to-Ground Search 23 Threshold Contrasts 23 Static Search «... 26 Motion and Visual Acuity 31 Motion and Search 35 Simulator Studies 37 Operational Studies 38

Mathematical Models 42 Summary: Solutions to Operational Problems 45

Acquisition» Detection» Recognition « 45 Usefulness of Data '. 46 Mathematical Models t7 Laboratory Tests (Simulators) '47 Field Tests 48

References *& Bibliography 55

Figures: 1. Maximum Angle of Depression for Pilot Looking Out of Aircraft

During Power Approach 2 2. Forward Area, Blind Range, Caused by Aircraft Obstruction 2 3. Blind Range for Various Altitudes and Maximum Depression Angle. 3 4. Percentage of Various Types of Terrain Seen Fro« Aircraft 4 5. Probability of Fairly Smooth Terrain Being in View 5 6. Probability of Moderately Rough Terrain Being in View 5 7. Probability of Rough Terrain Being in View 6 8. Angular Velocity Geometry for Air-to-Ground Search 8 9. Nomograph for Computing Angular Rate in Level Flight 11

10. Angular-Rate Contours of Terrain as Seen by Pilot During Level Flight at 50-Foot Altitude and 350-Knots Velocity 12

iii

[AVWEPS REPORT 8617

11. Angular-Rate Contours of Terrain as Seen by Pilot During Level Flight at lOO-Foot Altitude and 350-Knots Velocity 12

12. Angular-Rate Contours of Terrain as Seen by Pilot During Level Flight at 500-Foot Altitude and 350-Knots Velocity 12

13. Boundary Between Increasing and Decreasing Angular Rate as S or H Increases 13

1U. Angular Velocity of Ground Points as Seen From an Aircraft in Level Flight 13

15. Angular Velocity of Ground Points as Seen From an Aircraft at 600-Foot Altitude 1*»

16. Target Angular Rate With Aircraft Velocity at 350 Knots in Level Flight 15

17. Target Angular Rate With Aircraft Velocity at tOO Knots in Lavel Flight 15

18. Target Angular Rate With Aircraft Velocity at U50 Knots in Level Flight 16

19. Target Angular Rate With Aircraft Velocity at 500 Knots in Level Flight 16

20. Velocity Discrimination Thresholds 18 21. Degree of Scintillation Versus Resolution Using Landolt

Broken-Ring Chart 20 22. Eye Sensitivity, Solar Radiation, and Reflectance of a Red

Surface 21 23. Reflectance of Desert Terrain 22 2»». Liminal Contrasts for Round Targets Brighter Than Their

Backgrounds < 2*« 25. Landolt Ring, Showing Proportions 2»* 26. Threshold Contrasts for Rectangles of Various Dimensions 25 27. Average Probability of Detecting Rectangular Targets Against

a White Background , 26 28. Example of Field to be Searched, From Experiments of Boynton

and Bush (Ref. 39) , 27 29. Search Performance, From Experiments of Boynton and Bush 27 30. Targets Used in Experiments by Baker, Morris, and Steedman

(Ref. «♦2) 28 31. Search Time Versus Search Area for Experiment by Baker,

Morris, and Steedman 29 32. Search Time as a Function of the Number of Signals and

Partitions 29 33. Search Time as a Function of Target-Pseudotarget Size and

Contrast Difference 30 3*4. Search Time Versus Peripheral Discriminability, From Smith

(Ref. 45) 31 35. Average Search Time on Static Displays. From Erickson 32 36. Visual Acuity as a Function of Angular Velocity of the Tar-

get 33 37. Results of Display Scale Experiments of Kraft and Hamilton... 33 38. Legibility Curves for Letters Subtending About 39 Minutes to

the Observer 3U

Iv

NAWEPS REPORT 8617

39. Cuaulatlve Probability fop Each Sp««d fop High-Density Field, Fpoa VilliaM and BOPOW (Ref. 57) 35

HO. Visual Search Pepfopnance in a Moving Structured Field 36 HI. Ppobability of Detection, FPOB Ref. 62 38 H2. Search Performance ''0 H3. Probability of Detecting Ground Targets FPOB Contour-Flying

Helicopters HO HH. Target Acquisition Ranges in Visual Alp-to-Gpound Seapch HI H5. Pepcent of Tapgets Recognized in Visual Aip-to-Gpound Seapch. H2 H6. Calculated Tapget Recognition Ppobability With 177 Degrees

Between Sun's Position and Line of Sight H3 H7. Calculated Search Pepfopnance Rating as a Function of Alti-

tude and Velocity fop a 50-Foot Object, Such as a House H3 H8. Accumulative Detection Ppobability Against a JS-III Tank in

an Area 600 by 600 Yards HH

- 3U ■«—■*

NAWEPS REPORT 6617 i

ACKNOWLEDGMENT

The author wishes to thank Roy Dale Cole, who designed the nomograph used in this report; and Czema A. Flanagan, of the Research Department, who verified mathematically the nonequivalency of angular rates at all points in the field.

vi

MAVWEPS REPORT 8617

INTRODUCTION

At present, one of the critical problems encountered in military operations is that of identifying ground targets from aircraft. The problem is more severe if the aircraft is flying low and fast. Search may be made either directly throi^h the canopy or on a television or optical screen whose sensor is mounted on the aircraft. In any case, the problem is one of identifying targets moving with respect to the observer.

The purpose of this report is to discuss the geometric, kinematic, physical, and human engineering aspects of air-to-ground search. The results of pertinent analytic and experimental investigations are presented and briefly discussed. Although the value of these investigations is as- sessed, no attempt has been made to perform a comprehensive evaluation of each of them.

GEOMETRY OF AIR-TO-GROUND SEARCH

Two factors that affect the detection probability of a ground target from an aircraft arc obstruction of the field of view by (1) the aircraft and (2) by the terrain surrounding the target. Both factors are geometric in nature and must be considered in determining the detection probability for any given situation.

AIRCRAFT OBSTRUCTION OF VIEW

The great increase in jet aircraft performance in recent years has been accompanied by a decrease in the ease of searching from such aircraft. The faster the aircraft and the sleeker the aerodynamic design, the poorer the view from the cockpit. This view limitation Imposed by the aircraft is a function of the aircraft construction and the flying conditions, since angle of attack changes with aircraft loading, altitude, and velocity. The maximum line of sight depression angle has been measured for several aircraft in power approaches and is shown in Fig. 1. Data taken from Ref. 1 were used to plot the curves. If the aircraft is assumed to have a smaller angle of attack during a search flight, the curves would be raised somewhat. An example is shown by the A-UB depression angle dead ahead. If the angle of attack is close to zero, the dead-ahead point of the curve would be raised 8° U2' to the point indicated by the dotted line in Fig. 1. Gener- ally, the maximum dead-ahead depression angle for such aircraft varies

MAVWEPS REPORT 8617

between 10 and 20 degrees. This aircraft obstruction results in a large area benea '. the aircraft, blind area (shown in Fig. 2)« which the pilot cannot see without maneuvering. For instance, in level flight at 1,000 feet altitude with a depression angle of lu degrees, nothing can be seen

40

39 h

T -T • A «e 8 DEC. 4 2 MIN 1 * i J 3 M 12 N

• r • * 11 • 1 0 H

o A 3* 0 M 3« M

a F • A 0 m • • M

40 30 20 10 0 10 20 30 40

ANGLE LEFT OR RIGHT OF DEAD AHEAD.DC6

FIG. 1. Maximum Angle of Depression for Pilot Looking Out of Aircraft During Power Approach. Pitch angles during approach are nose up except for the A-3A (a two-place aircraft), which is noae down. The obstructive effects of some cockpit structures, gunsights, etc., although in some cases appreciable, are not shown in the curves •

y

///////

FIG. 2. Forward Area, Blind Range, Caused by Aircraft Obstruction.

MAVWEPS REPORT 8617

d««d «head on th« ground unless it is more than UvOOO feet ground-range ahead of the aircraft. Values of blind range can be taken fro« Fig. 3 for various depression angles and altitudes. This large blind area is one of the reasons pilots prefer to search through the side of the canopy, banking the aircraft during search. "S" turns are used often by the Navy (Ref. 2) and the Air Force when searching the ground for targets.

40,000

24000

i\l\ I \

luv SN k luv \ s 10,000 UV \ \ s 3,00c ii^m

l-V ^— -=K ►—- ■•■■

8,000 kA -*s «— — —

6,000 ll \ \ \Ä s. \\ k\ s, ^N

4.000 \\ Vs s,^ ^JUWJ w J t ' \\ \ \,

^^ X, ^j

2 2 000 \ v \, N 600 kooo '—

£ «,wu \ .\

> s.. V. — O

i 1,000 \ ■ \

^•s V 400 "^i \ ^

BOO \ v_ ^^H \ ^

SOO \fc ^^ \«

V

1 400 —>(

^^ fOO V v,^

2)0

*x ^100 "^H

100 ^

S 12 16 20 24 28 DEPRESSION ANGLE. DEC

32 36 40

FIG. 3. Blind Range for Various Altitudes and Maximum Depression Angles.

TERRAIN OBSTRUCTION OF VIEW

The degree of obstruction by the terrain has been estimated from contour maps by a graphical method (Ref. 3). The results can be presented in terms of percent of area within view (useful for estimates of reconnaissance effectiveness), or probability that a spot on the terrain

MAVWEPS REPORT 8617

at various ranges is within view (useful for weapon-de 11 very analyses). Figure t, taken from Ref. 3, shows the average percentage of various types of terrain within view from various altitudes. As the terrain becomes rougher, less of it is within the view of the air observer. Hnrever, an inversion of this trend occurs when going from rough to very r>ugh terraii It is hypothesized that the sides of the hills are less obscured when they are very steep; the associated vertical development leads to greater visibility. Figures 5, 6, and 7 show the probability that a spot m the terrain is within view from aircraft at various altitudes, as a function of ground-range. Since these results do not include the masking effects of foliage the results are expected to be overestimations of area within view and of probability of seeing a target. This was shown to be true in some cases where the method of obstruction estimation was ompared to the results of a field study1. These curves can therefore be wised to set the upper limits to the measures of obstruction.

100

1,000 1,500 2,000 2,900 ALTITUDE ABOVE TERRAIN, FT

3,000

FIG. «♦. Percentage of Various Types of Terrain Seen from Aircraft.

U. £. Naval Ordnance Test Station. Ground Target Visibility, by Carol Gill. May 1962. (IDP-1U87), UNCLASSIFIED.

Terrain Effects Upon Air-to- China Lake, Calif., NOTS, 15

NAVWEPS REPORT 8617

FIG. 5. Probability of Fairly Smooth Terrain Be- ing in View. Aircraft alti tude above terrain is shown on curves, in feet.

CMOUMD MAMGC, OF FT

t 4 • • 10 OROUND RANGE. THOUSANDS Of FT

12

FIG. 6. Probability of Moder- ately Rough Terrain Being in View. Aircraft altitude above terrain is shown on curves, in feet.

NAVWEPS REPORT 8617

Th« curves presented in this section can be used to estimate the probability of the target's being in view. Whether the target is detected and identified once within view is another problem which will be discussed in the section on Psychophysics of Air-to-Ground Search.

2 4 • I U GNOUNO RANGE. THOUSANDS OF FT

FIG. 7. Probability of Rough Terrain Being in View. Aircraft altitude above terrain is shown on curves, in feet.

KINEMATICS OF AIR-TO-GROUND SEARCH

The motion of the field being searched with respect to the observer can affect, in one way or another, search performance. In some cases, notion might increase performance by causing the observer to employ a more systematic method of search than that used in a static field. How- ever, search time limitation imposed by motion of the field tends to reduce performance, and under some conditions, motion per se of the field being searched will reduce search effectiveness.

6

NAVWEPS REPORT 8617

ANGULAR RATE OP THE GROUND

An equation giving the angular rate of ar.y ground object within view can be derived from the geometry shown in Pig. 8. The aircraft is diving in the plane ABCD with a dive angle « at a velocity V. The target. T, is lying In a flat plane beneath the aircraft. Its coordinates, Measured with respect to the aircraft, are H, distance beneath the aircraft; R, distance in front of the aircraft i and S, offset distance to the side (perpendicular to ABCD). The angular velocity of the point T can be measured with respect to the velocity vector V. The rate of change of a, the angle between V and AT, is the angular velocity of point T. It is seen that

.VH5^ AT »VH ♦ R ♦ S' (1)

and

ET s Vs2 ♦ (H - R tan 6)2

Then, by the law of cosines.

-♦■

(R + H tan 6 ) (H |£ ♦ R ^) cos 6

Prom Eq. <♦ one finds that

(3)

(R ♦ H tan 6) cos 6 /.iX cos a * —; i (»O VH2 ♦ R2 ♦ S2

Differentiating Eq. U with respect to time, one obtains

da =. (H2 + R2 * S2) (f + | tan *) cos 6 ^ dt (H2 ♦ R2 ♦ S2)3/2 »in a

(H2 ♦ R2 ♦ S2)3/2 sin a ^ ' (5)

. /(R sin 6 - H cos 6 )2 ♦ S2 /e. sin a « ./ (6) V H2 ♦ R2 f S2

NAVWEPS REPORT 8617

On« also has the relation:

tan 6 -

and

dH dt

IE dt

(7)

V = -

dR dt

cos 6 (8)

By substituting Eq. 6, 7, and 8 into Eq. 5 and simplifying the result,

one obtains

da . Vs2 ♦ (H cos g - R sin 6)2 „ dt 2 2 2

H ♦ R^ ♦ S (9)

FIG. 8. Angular Velocity Geometry for Air-to- Ground Search. AB, BF, and BD form an ortho- gonal coordinate system.

Equation 9 can be used to compute the angular velocity of any point on the ground beneath an aircraft diving or flying level with a velocity V. Values of the angular rate encountered in level flight can be taken from the nomograph shown in Fig. 9 for selected ranges of altitude and velocity. Some angular-rate contours for level flight over flat terrain are shown in Fig. 10, 11, and 12. The blind areas shown in the figures are calculated from a maximum depression angle of 16 degrees dead ahead and a maximum depression angle of 32 degrees at an angle U5 degrees back from dead ahead. It can be seen that at the lower altitudes of 50 and 100 feet, fairly high angular rates are encountered beyond the blind range, whereas from a 500-foot altitude the visible part of the terrain dead a- head is moving more slowly with respect to the pilot.

8

NAVWEPS REPORT 8617

The peculiar shapes of the 1 so-angular rate curves shown In Fig. 10, 11, and 12 cannot be easily understood by inspection of Eq. 9. Simplify- ing Eq. 9 to level flight (6-0) and differentiating with respect to S, one obtains

(

do\ CR2 - (H^ ♦ S2)] V

**' (H2 ♦ R2 + S2)2 (S2 ♦ H2)1/2 S dS . (10)

It is seen that for all points where R2 > (H fS),an increase in S will cause an increase in the angular rate. When (H2 + S2) > R2, the angular rate decreases with increasing S. It can be shown that the effect upon the angular rate of a point on the ground produced by a change in aircraft altitude is given by Eq. 10 when HdH is substituted for SdS. Fig- ure 13 shows the boundary on the ground where this change in sign of d(da/dt)is found. Data shown in this figure apply only to level flight, although a similar boundary between increasing and decreasing do/dt can be mapped for aircraft diving or climbing.

EQUIVALENT ANGULAR RATES

The angular rate of a point in the field encountered at some given altitude, velocity, and dive angle can be duplicated by flight at some other altitude, velocity, and dive angle. Such duplication of the angular rate of all the points in the field is not possible, however. Figure 1U shows the angular rate of ground points as seen from two different altitudes and velocities, selected so that at 3,200-foot ground-range dead ahead, the rates are equal. The peripheral field is moving faster at the lower than at the higher altitude. In Fig. IS, rates encountered in level flight are compared to those encountered in a dive. Again, velocity is chosen so that the rates are equal 2,U00 feet ahead of the aircraft. Even the peripheral rates are about equal at this range, and at shorter ranges the differences between rates are not large. At about 3,U00 feet ahead of the diving aircraft, the angular rate of the ground in zero; at points beyond that, the rate is negative. This is the main difference between dives and level flight. In level flight, the rate has the sane sign every- where and is zero only at infinity. A more detailed analysis of the angular rates encountered in dives, with application to landing aircraft and pilot judgments of motion, is given in Ref. U.

Figures 1U and 15 illustrate the differences in angular rates and hence, a possible source of error introduced in extrapolating search data from one flight condition to another. Such an error would be due to changes in the performance of human searchers that is induced by the different motions in the field being searched.

MAVWEPS REPORT 8617

INSTRUCTIONS FOR USE OF NOMOGRAPH

Giv*n:

Desired:

(1) S offset distance to target (2) H aircraft altitude (3) R range ahead to target (U) V velocity of aircraft

5?- • the angular rate of a point on the ground at (R, S) at

EXAMPLE

S - 265 feet H * 200 feet R s 2,000 feet V * 1,000 ft/sec

Procedure: (I) Find the desired S (265 ft) on the S-scale on the upper left of the nomograph.

(2) Then go across horizontally until the desired H-curve (200 ft) is intersected.

(3) Go straight down from this intersection to the index line V-WT

(»♦) Follow the curved lines down until the desired vertical range line (2,000 ft) is intersected.

(5) Draw a line from the index point on the lower right, through the intersection point obtained in (u), up to the index line X-Y.

(6) Now draw a line from this intersection point on the index line, through the desired velocity (1,000 ft/sec), across to the angular rate scale on the far left. The answer (4.67 deg/sec) is obtained on this scale.

(7) For convenience, two different scales, A and B, can be used. In the example mentioned, the A velocity scale is used, so the angular rate must be taken off the A do/dt scale. If the B scale had been used (V = 200 ft/sec) the corresponding angular rate would be 0.93 deg/sec, taken off the B scale.

Scales are given belov for conversion from ft/sec to knots and from deg/sec to radians/sec.

0 1

1 1

2 i

9 4 1

9 i

6 i

7 I

e i

9 i

to 1

OEO/SEC

0 I I

0.01 T !

0 04 i '

o.ot r 1 1

oot 010 1

0.12 1

0.14 T

01« 1 »*D/$EC

0 1

1 1

2 5 i

4 i

9 i

• 7 i

• • 10 n/SEC r 0 i

? t

I J 4

i 6 « KNOTS

10

BLANK PAGE

, «ÄMÜK»-'

2,ooo:

FIG. 9. Nomograph for Computing Angular

NAVWEPS REPORT 8617

2,000

H, ALT, FT

:or Computing Angular Rate in Level Flight a

INDEX POINT

11

NAVWEPS REPORT 8617

Nttt *ö

FIG. 10. Angular-Rate Contours of Terrain as Seen by Pilot During Level Flight at 50-Foot Altitude and 350-Knots Velocity. Angular rate is given on the contours in deg/sec. Blind area is also shown (darker portion).

XRfc

FIG. 11. Angular-Rate Contours of Terrain as Seen by Pilot During Level Flight at 100-Foot Altitude and 350-Knots Velocity. Angular rate is given on the contours in deg/sec.

*.ooo £

ooo k

O I.OOO ijOOO 3.000 4J0O0 5.000 «0OO r.ooo ».OOO 9fiOO tOfiOO //.ooo

FIG. 12. Angular-Rate Contours of Terrain as Seen by Pilot During Level Flight at 500-Foot Altitude and 350-Knots Velocity. Angular rate is given on the contours in deg/sec.

12

-^ ■"'*-■-' iHrniOMr*''"!

NAVWEPS REPORT 8617

n ~ OECREASes WITH INCREASING S OR H OR BOTH ot

I I If INCREASES WITH INCREASING S OR H OR BOTH 1 ' dt

FIG. 13. Boundary Between Increasing and Decreasing Angular Rate as S or H Increases. Aircraft is in level flight.

200

10.0 9.0 8.0

7.0 o UJ 6.0 «n

5.0 Ul o >-" 4.0 »- o o Ui ?0 > K t -1 3 «a z 2.0 4

i.ol 200 FT ALTITUDE, 1176 FT/SEC VELOCITY — 4OOFTALT'TU0El600FT/SEC VELOCITY i i i I I I I ! i_

500 400 300 200 100 0 100 200 300 400 500 S,OFFSET DISTANCE, FT

FIG. 1U. Angular Velocity of Ground Points as Seen From an Aircraft in Level Flight. Range to target is shown on the curves. (See Fig. 8.)

13

NAVWEPS REPORT 8617

200

i/>

10-0 9.0 S.0

7,0

6.0

*• S.0 g ► 4.0

9.0

2.0

—1 1 1 1 1 | r 400 FT/SEC VELOCITY, LEVEL FLIGHT 870 FT/SEC VELOCITY, 10-OEG DIVE

1.0

900 400 300 200 100 0 100 200 900 400 900

8, OFFSET DISTANCE. FT

FIG. 15. Angular Velocity of Ground Points as Seen from Aircraft at GOO-Foot Altitude.

TARGET ANGULAR RATE

Thus far, the discussion has treated the angular rate of part of the visual field and Is applicable to search considerations. Once a target has been detected during this search, the next step may be to make a pass over it and release or fire a weapon. It Is therefore advisable to consider the angular rate of a target as the aircraft files over It. When diving toward the target or flying at high altitudes, the target's angular velocity Is low (at or before weapon release) and has small effect upon search effectiveness and tracking accuracy.

Since during level flight at low altitudes the velocity effect could be a major one, curves are shown only for low altitudes and higher velocities. The target angular rates are shown in Fig. 16 through 1° as a function of aircraft altitude, range to target, and time to target. Tht<

IH

NAWEPS REPORT 8617

release points of a weapon in a vacuum and of a high-drag boirib ore also shown on the curves and will be referred to later. It is interesting to note for the cases shown that the angular rate of the target is fairly low and does not change rapidly at ranges over 5,000 feet. At ranges under 5»000 feet, when the aircraft is getting close to the target, the angular rate begins to increase rapidly. There may therefore be greater errors when tracking during low-level fast flights than during other weapon delivery modes. The angular rates described here, incidentally, are the same as those encountered by the ground observer attempting to track an aircraft.

«0

g.o

gTO

J 6 0 (- 4 ■so

> * JO

Ut

< 1 0

20 a 16

riMe

14

TO TARGET

12 10

, sec 8 6 4 2 .._2

' 11 L

—\ f '

HUM POWT •Mat «a««

Mem *•• i

J

— --. to»

m m

, .

- Mi i lit

vacuuf P

h u M M / j i • - —

r T -

Bi m dfii g£> r 0

•2,1 X» 10. Mb ■,c XXJ 6J( XX) 4, DOC Zi XX) 0

RANGE TO TARGET, fT

FIG. 16. Target Angular Rate With Air- craft Velocity at 350 Knots in Level Flight. Altitudes are shown on curves.

.« •• • • 4

TIME TO TÄUOCT.

It 10 •

MC • 4 t 0

ttli) r r _ «it. [AM * «mi Hi Um MM m IO«

Sao i i m

m 5 »o —

^

mi « 2Q m

1 o «A CWUMAUMM. _,

. »omt i -^ ^^ a

aa SP^ ^

0 HI HI Bi m^ s- =—■ - . 12,000 10,000 «POO 6^00 4£00

RANGE TO TARGET, FT

2000

FIG. 17. Target Angular Rate With Air- craft Velocity at UOO Knots in Level Flight. Altitudes are shown on curves.

15

MAVWEPS REPORT 8617

• 0

..BO

1 14 TIME TO

12 10 TMCCT.SEC

• • 4 2 0 " I T I n w

M ^ 7 O

1 < • iUMI »OHiT Uaa aati« UK KM _ 1

8 . «0

M

2 so

TTja f

-• 40 III 3 ,0

«* S « so

i

mi m ÖSy«/ 1 1 1 -^ %\ 0 ' * _'

Edg ^^- F J 1

^ ^ Ü X

0 Hi 11,000 tOJOOO •fiOO %flOO 4,000 i,000 0

RANOC TOTANMT.FT

FIG. 18. Target Angular Rate With Aircraft Velocity at U50 Knots in Level Flight. Alti- tudes are shown on the curves.

12,000 lOjOOO •,000 tfiOO 4,000

KANOZ TOTANSET, fT

1,000

FIG. 19. Target Angular Rate With Aircraft Velocity at 500 Knots in Level Flight. Alti- tudes are shown on the curves.

16

NAVWCPS REPORT 8617

MOVING TARGETS

It has been assumed In the preceding discussion that the target Is stationary with respect to the ground. Motion of the target can enhance detection In three ways: (1) a new target. Is created by the notion, such as the wake of a ship or a dust cloud behind a tank; (2) the change In location of the target due to Its motion is noted; and (3) in some cases, the motion per se of the target attracts the observer's eyes. Although the last two factors are difficult to assess, a consideration of theory and experiment will illustrate their relevance.

Consider two objects moving parallel through the visual field with angular velocities w, and «2♦ A differential threshold for angular velocity may be defined as

Aw s w. — w. (11)

That is, the difference between the angular velocities of the two objects must be at least Aw or an observer cannot with any confidence tell that there is a difference.

Laboratory measurements of Aw have been made with moving spots on an oscilloscope, rotating disks, needle pointers, and other such devices. It has been found that Aw is a function of the angular velocity of the reference object such that .

W = ^ü = constant (w * 0) (12) w

within certain limits. W is known as the Weber ratio. From data summarized in Ref. 5 (Fig. 20), it is seen that W = 0.1U for curve 1 and W = 0.08 for curve 2.

Consider the simple case of a moving target being viewed from an aircraft flying level with constant velocity. If the target is moving along the ground track of the aircraft, its angular velocity would be

= . (v^vW ^ („,

as compared to the angular velocity of points on the ground about the target, which is given by

HV w R H2 ♦ R2

(1U)

where v is the velocity of the target and the other symbols have been defined earlier. It can be shown that

W = £ (15)

If a Weber ritio could be determined for the above situation and were found to be o.io, say, it could be concluded that targets moving less than one-tenth the aircraft's velocity would not be spotted by virtue of their motion per se .

17

NAWEP8 REPORT 8617

10 o

%

s

o 010

G 3 M >

3

5 ooil-

0 001 001 0.10 1.0 10.0 100

ANGULAR VCLOClTY. OEG/SCC

1. Adjacent stimuli 2. Separate stimuli

3. Superimposed stimuli, 3.6 deg field 4. Superimposed stimuli, 15.0 deg field

FIG. 20. Velocity Discrimination Thresholds.

OPERATIONAL EFFECTS

It should be pointed out that two additional factors must be considered in some cases when applying the angular rate information presented here. The angular rates are measured with respect to the aircraft 's ground velocity vector, and in the presence of crosswind, the velocity vector and the aircraft reference line are not coincidental. Also, since it is not always necessary to keep the eyes directed along the velocity vector, the searching pilot can greatly reduce the angular rates simply by tracking the target with head and eye movements. The same results could be obtained with a properly designed search or bombing system.

18

MAVWEPS REPORT 8617

OPTICAL ASPECTS OP AIR-TO-GROUND SEARCH

Detection involves light entering and passing through the atmosphere, reflecting off the target and its surroundings, and again passing through the atmosphere to be received by the eye or a sensor. The amount, spectrua, and direction of the emitted light, the optical properties of the atmosphere, and the reflectance of the target and its surroundings affect target detection probability and target detection ranges .

CLOUDS

Clouds have two general effects upon target visibility: obstruction of the target, and diffusion of the light coming from the sun, thus affecting the way the target is illuminated. In some operation analyses, an estimate of the amount of cloud cover is useful. Such data can be found in Ref. 6 through 9. These references give average monthly values of cloud cover and are averaged into seasonal values in Ref. 10. Similar information can be obtained from Ref. 11, 12, and 13.

Illumination, in foot-candles, falling on a fully exposed horizontal plane at any point on the earth at any hour of the day or night, can he found from the charts presented in Ref. 1U. In this reference it is stated that when the sun is obstructed by thin clouds, the value of illumination should be divided by two: for average cloud conditions obstructing the sun's rays, the values given for clear days should be divided by three; and for dark stratus clouds, the values should be divided by ten.

In some mathematical egressions for visibility parameters, the brightness of the sky or the ratio of the brightness of the sky to the brightness of the target background appears. Such data can be found in Table 5.2 of Ref. 15, Table 3 of Ref. 1€, and Table C of Ref. 17.

In experiments concerning visual resolution, it was found that resolution will increase appreciably as skies approach overcast conditions (Ref. 18) The cloud cover reduces the effects of "shimmer" (see next section).

ATMOSPHERIC ATTENUATION

Light passing through the atmosphere is subject to absorption and scattering; thin reduction in intensity of a beam, called attenuation or extinction, has been measured under a variety of conditions. The attenuation coefficient, a, ?s defined by

I = I0 e"ax (16)

where I and I are the intensities of a collimaled beam of light entering and emerging from a layer of air x-units thick. Values for a can be found

19

NAVWCPS REPORT 8617

or calculated from data given in Ref. 15 through 21 and generally range fron 0.05 to 0.5 per kilometer. Such information can be used in calculations of target-back ground contrast attenuation for application in visibility models used to compute detection and recognition ranges. Light scattering of particles in the atmosphere results in a luminance of the "space" between the observer and the target. This path luminance acts to reduce target-background contrast and hence make the target more difficult to find.

The irregular refraction by the atmosphere of the light rays reflec- teo by an object frequently makes identification of the object difficult The index of refraction of the air sometimes varies rapidly and irregu- larly from place to place, causing shimmer (Ref. 15, 18, 22 and 23). An example of the degradation in visual resolution caused by shimmer is shown in Fig. 21 (taken from Ref. 22). Such degradation can be assumed for some conditions in recognition range calculations.

« x -

m

> >-

8 5-

o

0 2 4 • • 10

SCINTILLATION (PERCENT MODULATION)

FIG. 21. Degree of Scintillation Versus Resolution Using Landolt Broken-Ring Chart

TERRAIN AND TARGET REFLECTANCE

One of the factors that determines whether a target can be detected is the contrast in brightness and color between the target and its background. A very broad classification of backgrounds by color is given in Ref. 2M. TN» role of color is indicated in Fig. 22, where the eye^

20

KAVWEPS REPORT 8617

sensitivity, the intensity of sunlight, and the reflectance factor of a red surface are shown as functions of wavelength. Reflectance data for various types of terrain are given in Ref. 25, 26, and 27. Figure 23a, taken from Ref. 27, shows the reflectance of various types of desert terrain. Figure 23b shows that the sun's angle makes a considerable difference in the reflectance, and hence brightness, of terrain. The data of Ref. 25 have been converted to a single reflectance factor by using values weighted according to the relative sensitivity of the human eye to dif- erent wavelengths. These values are shown in Table 1 as taken from Ref. 28, page 10.

VIOLET iLUC OREEN VEUOMT RED 100

0.S0 -

> 0.60-

«/)

3 S

0.20-

400 600 600 700 WAVELENGTH. MILLIMICRONS

FIG. 22. Eye Sensitivity, Solar Radia- tion, and Reflectance of a Red Surface.

The reflectance of typical ground targets has not been found in the literature. A single value for reflectance does not exist in most cases: each man, vehicle, or bridge has a number of different reflectances. Some examples of average contrast of a target against a background are available however. (Contrast is defined in the next section.) From measurements on models reported in Ref. 29 it was found that olive drab troops and vehicles in California desert have a contrast of about 0.6. Olive drab troops and vehicles on black pavement, some dirt roads, and green grassy fields, have a contrast of about 0.3.

21

NAWEPS REPORT '8617

(a)

400 460 SOG S50 «00

WAVFLCNOTH. MILLIMICRONS

•50

BRUSH AND WIND-»LOWN tANU

(b)

SO 40 SO 60 TO «0

ZENIIH DISTANCE Of SUN, DEO

FIG. 23. Reflectance of Desert Terrain.

TABLE 1. Reflectance Factors for Various Surfaces

Surface Weighted reflectance factor

Black earth 0.03 Earth roads 0.03 Paved roads 0.09 Buildings 0.09 Forest (winter) 0.03 Forest (autumn) 0.16 Forest (summer) 0.10 Grass fields o.oa Dry meadows 0.08 Lush grass 0.10 Fresh snow 0.77 Open sea 0.50

22

NAVWEPS REPORT 8617

PSYCHOPHYSICS OF AIR-TO-GROUND SEARCH

The final step In the visual detection process is the reception of light rays from the environment by the observer, and the selection and interpretation of enough pertinent information to permit target detection. The physiological and psychological characteristics of the human visual system are therefore important factors in the estimation of detection ranges and probabilities of detection. Several experiments that have been made in the laboratory and are described briefly below furnish data on visual capabilities.

THRESHOLD CONTRASTS

The brightness of a target and its background can be calculated from illumination and reflectance information or can be measured directly. The contrast of the target is defined as

c = -V-^ (17) B

b

where Bt is the brightness of the target and Bb is the brightness of the background. For targets darker than their backgrounds the contrast is negative, but since it has been shown that targets are equally visible if their contrasts are numerically equal, the sign need not be regarded as so important. It must be noted that contrast due to color differences is not included in Eq. 17. Color contrast is more difficult to access ana- lytically, and the absence of large color differences in military situations (tanks are seldom painted blue) has led to little investigation of this factor.

In experiments conducted during World War II, threshold (or llminal) contrasts of human observers were determined for various target sizes and adaptation brightness. Llminal contrast is simply defined as that value at which detection occurs 50% of the time. A spot of light was projected onto a white screen located 60 feet from the observer who indicated where the spot appeared. A large amount of data were analyzed and plotted as shown in Fig. 2U that was taken from Ref. 30.

A later experiment by Blackwell and Holdauer (Ref. 31) determined the threshold for various positions of the target on the retina. Targets subtending 1 minute of arc to the observer and presented for 0.01 second were used.

The data of Blackwell reported in Ref. 30 have been extended by Taylor in Ref. 32 and 33. Experiments similar to those described in Ref. 31, but thought more useful in visual search calculations, are reported in Ref. 32. The targets were exposed for 0.33 second instead of the 0.01 second in Blackweli's experiment.

23

NAVWEPS REPORT 8617

Visual contrast thresholds wer« determined in Ref. 35 in a manner differing fro« Blackwell's and Taylor's: Landolt rings were used as targets (Fig. 25). The thresholds were consistently larger than those found by Blackwell.

I i i 8

.9 -4 .1 -2 .1 0 12 3 LOG ADAPTATION SmGHTNESS, FT-L

PIG. 21*. Liminal Contrasts for Round Targets Brighter Than Their Backgrounds. The target was presented in only one position for a sufficient time to attain maximum frequency of correct response. The liminal contrasts are for 50% correct identification (above chance).

1 QK r ^fM

i

W^ M^v

i i I L !77 r ■ — ^•fe'^Hh m^-'y i

! i !^ ̂ y FIG. 25. Landolt Ring, Show- ing Proportions.

Contrast thresholds were measured for rectangles with length-width ratios of 2 to 200 in a study described in Ref. 36. For target areas less than 100 square minutes, square targets had lower thresholds than rectangular ones: the greater the ratio of length to width, the higher the threshold (Fig. 26). The data are further analyzed in Ref. 37 from which

2«»

NAVWEPS REPORT 8617

Fig. 27 was obtained. It is seen that for 100% detection the contrast must be about two times the threshold contrast. Th« detection probability may be even lower than that shown in Fig. 27. According to Ref. 16, page 9U( "the probability of an observer voluntarily reporting the presence of a liminally visible.target is nearly ztro". Threshold contrasts have also been found for other forms. Crosses, for example, were used in an experiment described in Ref. 38.

0.01 J 10 30 ICO 300 1,000

ANGULAR AREA, SO MIN

FIG. 26. Threshold Contrasts for Rec- tangles of Various Dimensions. A 3-second monocular exposure was used with a brightness of 2,950 foot-lamberts.

In the above studies, backgrounds of uniform luminance were used. In a study reported by Bixel and Blackwell (Ref. 39), threshold contrasts were found for circular targets as viewed against a background made up of ball bearings painted gray. A "regular pattern of luminance nonuniformity" was produced. It was found that when the contrast of the target was ex- pressed with respect to the brightness of the immediately adjacent area and not with respect to the average brightness of the whole background, the results were the same as those obtained with the uniform background. "This implies that visibility thresholds are determined by target contrast at target borders rather than by some kind of average contrast."

25

NAVWEPS REPORT 6617

z o o Ui

UJ o

i 02

20

CONTRAST RATIO ( &£IiJit= CONTRAST \ VTMRCSHOLO CONTRAST )

FIG. 27. Average Probability of De- tecting Rectangular Target* Against a White Background,

STATIC SEARCH

The experiments described above included very little, if any searching for targets« and it was not necessary to discriminate target fron nontarget objects. In searching for ground targets from aircraft, however, it is often necessary to search over a fairly large area, and to make judgments on a number of objects. A number of laboratory experiments have been carried out requiring this kind of search.

In the study of Boynton and Bush, observers were asked to search for a specified target located among «, number of irrelevant forms (Ref. UO). The target and objects were located on a circular, back-illuminated glass plate as shown in Fig. 28. Typical results are shown in Fig. 29, taken from Ref. Ul and U2. Performance decreases as search time decreases, and as the number of objects in the display increases.

Ir. the study described in Ref. U3, observers were asked to search a circular field for a target (see Fig. 30) located among a numl er tf similar objects. Search time increased with search area size (see Fig. 31). Since

26

NAVWEPS REPORT 8617

FIG. 28. Example of Field to be Searched, From Experiments of Boynton and Bush (Ref. 40)

EXPOSURE TIME, SEC

FIG. 29. Search Performance, From Experi- ments of Boynton and Bust . Object density is shown on each curve.

27

NAVWEPS REPORT 8617

FIG. 30. Targets Used In Experiments by Baker, Morris, and Steedman (Ref. ^3). The figure above is not representative of the displays used, but only of the objects in the displays.

the number of forms on a search area was rougly proportional to the search area size, it was felt that the primary factor in the increase in search tine was the increase in the number of irrelevant forms.

In one experiment a square white field was partitioned into equal sections by black lines (Ref. UU). This field contained squares, diamonds, and triangles, the latter being the targets. Search time increased as the number of nontarget objects and partitions of the field increased. The results are shown in Fig. 32.

A circular display containing many small dots (circular pseudotar^ets) and one designated target (a square, triangle, hexagon, or pentagon) was used in a search experiment by Smith (Ref. U5). Typical results (for the square target) are shown in Fig. 33. It was also found that the triangle was the easiest target to find, and then in increasingly diffi alt order, a square, a pentagon, and a hexagon. In further experiments, peripheral discriminability of the targets was meaaured and compared to search time for that target (Ref. 46 and Fig. 3U). It is seen that the easier it is to discriminate the target peripherally, the quicker it can be found in a display containing other objects.

28

NAVWEPS REPORT 8617

FIG. 31. Search Time Versus Search Area for Experiment by Baker, Morris, and Steedman. 12 it

SEARCH AREA, SO. IN. 24

20 40 60 80 NUMBER OF IRRELEVANT SIGNALS IN DISPLAY

FIG. 32. Search Time as a Function of the Number of Signals and Partitions. Num- ber of partitions in display are shown on each curve.

29

NAVWEPS REPORT 8617

10.0

• 0

«0

J 1

40

10

1.0

OJ

o«

Q4 -

r 1 25« CIRCULAR PSCU00TAR6ETS,

I SQUARE TARQCT

«RCENT CONTRAST OIFFERCNCE

X ± -L 10 20 30 40

PERCENT SIZE DIFFERENCE

80

FIG* 33. Search Time as a Function of Target-Pseudotarget Size and Contrast Difference.

In three experiments described fully in Ref. U7t a peripheral visual acuity score was obtained for 16 subjects using a Landolt ring as the target. The time it took these subjects to find a target in a display containing other objects was also measured. Those subjects with the higher peripheral acuity (PA) scores tended to find the target quicker than those whose PA scores were lower, as was suggested by tests conducted by Smith, Boynton et al. Two types of displays were used: one contained "blobs" and the other rings. Search time was longer and less affected by object density for the blob displays than for the rings. The search task was repeated using only ring displays as part of the second experiment, and the average search times were much the same as those measured in the first. The third experiment employed a linear cue in ring displays by adding a black line to the display. The target was located somewhere along this black line. Search time decreased greatly and was not as affected by object density as it was in the displays without the cue. These search timei are shown in Fig. 35.

Other studies of static search are included in Ref. U8 where such parameters as blur, eye movements, vigilance, and strategy are discussed.

30

NAVWEPS REPORT 8617

X u.

(9

K 141 I a c a.

TARGET ■ PSEU0OTAR6ETS o O •

DISPLAY CONDITION

FIG. 34. Search Time Versus Peripheral Discriminability, From Smith (Ref. U6). Peripheral location in the figure is the distance from the point of fixation at which the target could be discriminated from a pseudotarget 75% of the time (50% corrected for chance).

MOTION AND VISUAL ACUITY

Ludvigh and Miller were the first to do considerabTs work with moving targets. A series of reports issued by the U. S. Naval School of Aviation Medicine, U. S. Naval Air Station, Pensacola, Florida, describe th98e studies. Some of the studies are also discussed in Ref. »»9 and 50. Figure 36 shows that monocular visual acuity as measured with a Landolt ring deteriorates as the velocity of the target increases.

A good discussion of the work of others in the field and a presentation of the results of an experiment conducted at Tufts University is given in Ref. 51. Landolt rings were used as in the Ludvigh and Miller studies to determine dynamic visual acuity. Although acuity deteriorated with increasing velocity, it was found that lengthening the tracking time

31

NAVUEPS REPORT 8617

29

2 4

22

20

I.S

I.«

3 1.4 in

1.2 -

1.0 -

oe

1 1 1

1 ^^^\ etoes ^^^ *\

^ * \ s • r ^ H ^^ ^ • •

r *" i I • i RINGS • «" \~ * H ^ ^ • I x r *^ H r U -1

RINGS WITH LINEAR CUC J-*»-^,*^4

o**0^*" \

r^— 1 r i 1

IS 32

OSJiCT DENSITY

FIG. 35. Average Search Time on Static Displays, From Erlckson.

or target exposure time, or both, slowed this deterioration. These conditions resulted In better acuity scores than In the Ludvlgh and Miller studies.

In a study by Burg and Hulbert (Ref. 52), the target used was the checkerboard pattern of the Bausch and Lomb Ortho-Rater. The average dynamic acuity score was much better than that reported by Ludvlgh and Miller. A different test object and binocular tracking over a longer distance may account for some of this difference.

In a similar study (Ref. 53), a Landolt ring projected upon a screen by a rotating projector was the moving target. The results were much the same as those reported above.

Display scale studies performed by Boeing Airplane Company (Ref. 5H) indicated that an optimum display scale and an optimum viewing time ex- isted for best target recognition. The curve is shown in Fig. 37.

32

NAVWEPS REPORT 8617

it 1 i r

i «

s

3

REF. 4i • GROUP o CROUP

GROUP ^ LARGEST THRESHOLDS, * REF. 50

o MALE SUBJECTS. REF. 51

±. J- J- J. -L X J^ tO 40 «O M 100 ItO 140 140

AN4ULAR VKLOCITY OF TAMIT, 014/SKC

25

20

90

M 15

FIG. 36. Visual Acuity as a Function of Angular Ve- locity.

ISO

VIEMMNOTMK.SCC

404 SJOZ 01%

to

2

FIG. 37. Results of Display Scale Experiment of Kraft and Hamilton.

, I / ^ / / / /

f

/ / / /

-J. / /

005 050 100 05JECT HCIOHT, IN

400

33

NAVWCPS REPORT 8617

The legibility of moving letters and numbers was measured and results of the studies are reported in Ref. 55 and 56. Legibility thresholds are shown in Fig. 38. The symbols were moving in a column from top to bottom. In a later experiment, the relative legibility of the same letters was determined for velocities of 22.5, 31/ , and 36.0 deg/sec past the observer's eyes. Legibility rankings at all velocities were found to be significantly correlated with each other at the 0.001 level. No significant correlation was found with the rankings and the frequency of occurrence of letters in the English language.

•o

S5

u

aj

so-

40

IS

10

s

1 r— VCWING (MSTANCC

I I St IM.

T.I STO DEVIATION

I »TO DfVIATIOM

-i_ -U -i- J_

to

to

70

to

«» IÜ >■■ t 5 *

SO K

80

40

to

10

M

t 10 It 20

APCRTURC HCItNT, IM.

25 SO

FIG. 38. Legibility Curves for Letters Subtending About 39 Minutes to the Observer.

After a review of research published in Ref. 57, it was concludvd that targets moving across the visual field can be equated to stationary targets flashed on for a single brief exposure, equal in duration to the time required for the target to move across a point on the retina.

3H

NAVWEPS REPORT 8617

MOTION AND SEARCH

Visual target detection from an aircraft was simulated by making ob- servations from moving automobiles in experiments reported in Ref. 29. Toy soldiers and vehicles were used as targets, and the geometry was equivalent to that encountered at a flight altitude of 100 feet. Reconnais- sance scores were not appreciably affected by the speed of the automobile when the course was run at simulated speeds of 1,200 mph or less. The amount of actual searching in these tests is not known.

An experiment was performed at the Minneapolis-Honeywell Company using moving typewritten capital letters (Ref. 58). The same letters were always present in the field, which moved but did not change. Search performance began to decrease at display speeds between 3 and 16 deg/sec. The probability of locating the target, which was always within view somewhere in the field, is shown in Fig. 39. Again, it is seen that the higher the speed, the lower the probability of detection.

FIG. 39. Cumulative Probability for Each Speed for High-Density Field, From Williams and Borow (Ref. 58).

Some of the variables involved in detection of moving targets were studied in the experiment described in Ref. U7. The task was to detect the target, a Landolt C, among a number of solid rings. The rings were

35

NAVWEPS REPORT 8617

glued to a long white opaque belt that was moved vertically past a square window. Search performance generally deteriorated with an increase in belt velocity, ring density, or both. The probability of locating a target was strongly dependent upon its position in the display: the closer it was to the vertical centerline of the display, the more often it was detected. It was tentatively concluded that motion per se had no detrimental effect upon searching fields that were moving up to 10 deg/sec. The percentages of targets found in "equivalent search time" in moving and static displays were very much the same. This suggests the possibility of an extension of the statement in the previous section taken from Ref. 57 concerning apparent equivalency of detecting moving and static targets. For the velocities investigated, probability of detection appears to be the same when »earch also is involved. Foveal acuity seemed to be a better prediction of search performance than did peripheral acuity at the higher field velocities. The probabiUty of finding the targets is shown in Fig. HO.

100

oto

(a>

8

t

CM

040

020 -

1 —1 T ■ 1 f 1

k "C"^--^ » k

* J

Si

-v^v X ^J ^

r-« « • _l 1— ) 0 1 4 • • 10 It

TAAOCT POSITlOM.OttTANCE LIFT OR RltMT Or VINTICAL CfMTKNLINI, IN.

(b)

*

►■ ►•

100

oao

o»o ■

040 ■

o.io -

FIG. UO. Visual Search Performance in a Moving Structured Field. Velo- city, in deg/sec, is shown on curves in (a) i object density is shown on curves in (b).

36

NAVWEPS REPORT 8617

SIMULATOR STUDIES

It appears that most laboratory studies yield data that are useful in answering academic questions rather than the operational questions encountered by the military. When no other information is available, such data are used in operational analyses; but if time, space, and the budget permit, it is preferable to gather data in a situation approxi- mating the one of interest. Although the boundary between "academic" laboratory experiments and simulations is arbitrary, the author considers the studies described above to be »ore of an academic nature than operational, with the exception of the work reported in Ref. 29 where toy soldiers and vehicles were used as targets .

The study reported in Ref. £3 closely simulates a photo interpreter's task and might be considered to simulate, to a lesser degree, search from a high-flying aircraft, say 10,000 feet. In this study there were narked nonuniformities of display coverage by the observer. It was also found that the time needed to locate critical objects was markedly increased by degradation of the display resolution. A number of other findings were made regarding eye movements during search and are reported in more detail in the sources listed in Ref. 59.

Studies carried out on contract at Hughes Aircraft Company closely simulated photo interpreting tasks in which performance was measured as a function of display resolution (Ref. 60). It was concluded that since en- larging the scale has a detrimental effect at poor resolutions and a bene- ficial effect at better resolutions, the most effective means of improving the operator's performance is to increase display resolution. In an experiment conducted by the same group (Ref. 61), strip maps we^e moved past a window through which an observer was to search and identify specified targets. A fairly low percentage of targets were recognized at both display velocities used.

In an exploratory study reported in Ref. 62, it was shown that there was no significant difference in the performance of photo interpreters using different (vertical or oblique) or additional (vertical and oblique) views of the target area.

Reference 63 describes a simulator study using a rural landscape model built to a scale of 1:108. Small model tanks, other vehicles, and foot soldiers were used as targets. The experiments were run at two illumination levels: 0.02-foot candle (that of a three-quarter moon), and 2.0-foot candles (mortar flare). The influence of the light source position upon visibility was determined, an4 the simulated ranges of detection and identification were obtained for the various targets. It was concluded that a single flare may be more effective than a combination of flares since multiple flares in certain locations will attenuate contrast, thus reducing visibility.

37

NAVWEPS REPORT 8617

A terrain model 10 feet an a side and built to a scale of 1:600 was used in the study reported in Ref. 6H. An actual site within a park containing trees» grass, planted fields, buildings, and a number of roads was copied in the model. The target used was a convoy of three vehicles parked along a road. Data collected on this simulator were compared to data collected over the actual park using actual vehicles as targets and nine Navy pilots in RC-H5J aircraft (at 130-knots airspeed) as observers. The probability of detecting the target both in the simulator and in the field is shown in Fig. Ul. It is seen that there is a large difference between the laboratory data and the data collected in the field. However, the internal relations of simulator and field data were enough alike so that data for the former could be used to extend the usefulness of data gathered In tne field.

7,000 14,000 ti.ooo SLANT RAMOC. FT

21.000 16,000

FIG. Ul. Probability of Detection, From Ref. 63

^RATIONAL STUDTES

If there is a need for data describing human capabilities in air-to- ground search, the most reliable approach is to arrange for a nurber of humans to fly around in airplanes and collect data on their search capabilities. All variables would be present, such «t time and attention

38

NAVWEPS REPORT 8617

necessary for actually flyinp, the airplanes, vibration of the aircraft, air turbulence, and optical distortions by the windscreen and atmosphere. In spite of the difficulty and expense in such an undertaking, a number of such studies have been made. The study reported in Ref. 6'* described in the last section is an example. Those data can be used for estimating the detection range of a three-vehicle convoy painted oattleship gray and parked along a road. Other conditions, such as the 130-knot speed of th? aircraft, must also be kept in mind.

A series of studies known as Project Longarm were conducted by George Washington University, under contract to the Army, and are summarized in Rof. 65. Targets such as riflemen, mortars, rocket launchers, jeeps, tanks, bunkers, and trucks were used in the tests. Forty-two observers were flown, one at a time, over the target area in an O-TA aircraft at about 92 mph and 350 feet above the ground. Targets smaller than 5 square milliradians were undetected by most observers; whereas targets subtending more than 50 square milliradians were detected if they w^re exposed for 5 seconds or more, were relatively unconcealed and were viewed under good visibility conditions. Targets were detected more frequently when they were moving than when static nary. It is further staled in Part I of Ref. 65: "Results from Projec 'igarm clearly indicate that, at low altitudes only (500-1,000 feet), ac r -. observer performance was lower at high speeds (275 knot;) than fcr low speeds (87 knots).11 This difference in performance almost disappeared at altitudes above 2,500 feet.

In a stur*y reported in Ref. 66, it was found that the method of searching as well as the velocity of the aircraft carrying the observer affected search performance. Twentv-four observers searched for targets ranging in size from a tank to an Automatic rifle. The observers were flown past the targets in an 0-1A aircraft and an 0H-23 helicopter at 200 feet above the ground, four different methods of search were used. The results, shown in Fig, U2, were typical: performance decreases with increasing velocity, and some search methods are better than others.

Tests of target recognition capabilities from helicopters flying at or below tree-top level (contour flight) at 60 mph are described in Ref. 67. Tanks, jeeps, self-propelled guns (Scorpion) , and mortar emplacements put in the field (variable terrain) in strategic defensive locations were amoTig the targets used. Thirty-two pilots with aerial observation training or experience acted as observers in the tests, which were run from 7:00 a.m. to 5:00 p.m. Part of the results taken from Ref. 67 are shown in Table 2 and ^ig. 43. The probability of detecting the targets is quite low, and the detection ranges are considerably shorter than many theoreti- cal analyses have predicted.

A survey of the literature summarized in Ref. 68 revealed that conditions of field tests that have been carrieö out are so varied that com- bining the data to obtain a single estimate of search capabilities is risky. Nevertheless, in the absence of other estimates, this was done.

39

NAVWEPS REPORT 8617

The results, which combine data collected from a number of altitudes, velocities, aircraft, targats, and terrains, are summarized in Fig. uu and U5. The width of the curves reflects the confidence one might have in them.

Additional operational data that were not available for the summary given in Ref. 68, can be found in Ref. 69 and 70.

•o

r»

7«

•■

t

•o -

06 -L

FIC. H2. Search Performance. Search methods are shown or curves.

oSl0EM0VE

.FOeWARO 0M0VC

O FORWARD FIX

OSIOC FIX

±. 40 TO 100

AIRCRAFT VELOCITY, MFH

1,000 1,000

GROUND RAN8E TO TARKT, FT

~no.40

IfiOO

FIG. U3. Probability of Detecting Ground Targets From Contour-Flying Helicopters.

Enough such data do exist to provide a basis for some operational analysis estimates. It often seems, however, that for specific problems data do not exist, and tests must be run or preliminary estimates made from the results of using laboratory data in mathematical models.

«♦0

MAVWEPS REPORT 8617

TABLE 3. Probabilities and Ranges of Target Types

Recoilless rifle Item Tank

Support platoon Scorpion

Machine gun

Probability of detection at or before 0-50 yard minimum range

0.39 0.15 0.13 0.12 0.06

Probability of correct identification

0.35 0.12 0.12 0.06 0.05

Maximum range of detection, yard

1,U00 950 700 1,350 620

Range at which greatest frequency of detections occurred, yard

300 50 300 100 100

.20,000 - z o 0 tu

o

w o z <

10,000 —

1 1 1

1

— —^

!

— .*■...

»,-■■■ -■■■ %:■:■:■■■'..

—|

.

1 L.

1 1 1 LARGE LARGE SMALL STATIONARY

PROMINENT STATIONARY STATIONARY TROOPS TARGETS VEHICLES VEHICLES

FIG. UH. Target Acquisition Ranges in Visual Air-to-Ground Search. As an example of the range of variables, arrows show the range of data collected at a particular point on each curve.

«♦1

NAVWCPS REPORT 8617

!00

w »o

s

J2 60 Ikl V 9

Ö

.)■■' lii

40

20 -

mi ;4'

MOVING MOVING STATIONARY OPEN CAMOUFLAGED

ARTILLERY VEHICLES VEHICLES TROOPS ARMOR

FIG. »»S. Percent of Targets Recognized in Visual Air-to-Ground Search. Arrows show the range of data collected at a particular point on each curve.

MATHEMATICAL MODELS

Alaost all of the Material presented thus far in this report has dealt with direct Measurements: cloud cover, atmospheric attenuation, terrain reflectance, and visual capabilities. If operational data similar to that discussed above are available, knowledge of such individual variables is unnecessary. When there are no applicable operational data, however, preliminary estimates are often made by calculating detection ranges according to certain assumptions regarding the search process. For exaaple, consider the simulator and field studies already described in Ref. 6H. Since the internal similarities of both were the same, a transformation factor was used to extend the range of the field-test results. This factor was determined between corresponding simulator and field data and then applied to values obtained on the simulator under conditions that had not been duplicated in the field. This operation, which appears to be direct and reliable, yields curves similar to those shown in Fig. H6.

«♦2

MAVWEPS REPORT 8617

1.0

0.8

ot 0.6

-< 04 -

82 ^a 0.21-

0 -

T t 1 1 r

ALTITUDE, FT

5700 7»00 >00

J. 0 9 10 15 20 25

SLANT RANGE, THOUSANDS OF FT

FIG . US . Calculated Target Recognl- ^ tion Probability With 177 Degrees

Between Sun's Position and Line of Sight.

(9 Z

g (E

Ul u <

« o FIG. U?. Calculated Search Per- formance Rating as a Function of Altitude and Velocity for a 50- Foot Object, Such as a House.

—1 '~*~' ——CL" 100 M^H

»000

r — — • OO MI»H J

100 /

-^

1 1 1 _„!. 1 • 5 10 15 20 29 30

ALTITUDE , THOUSANDS OF FT

Figure U7 shows the results of a process a step further from reality, where relative search performance was calculated from the results of laboratory experiments described in Ref. HO, Ul, and H2 (see Static Search section). It was assumed that trends in search performance in a display similar to that shown in Fig. 27 are the same as trends encountered in the field. An optimum altitude of search for a given aircraft velocity and relative search performance at that altitude and velocity can be derived from the figure. It is suggested in Ref. H2 that the range of field data could be extended by using this model, as was done in the field and simulator studies described above.

The personnel at Scripps Visibility Laboratory also use the results of experiments performed specifically to gather data for a mathematical model. Detection and recognition ranges of targets are calculated from inputs such as atmospheric attenuation data, target-background contrasts measured with either a small model or the actual target, sun's angle, visual thresholds measured in the laboratory, and others. An example

U3

NAVWEPS REPORT 8617

of such a study is given in Fef. 71, where the detection and recognition ranges of a tank and a radar van are calculated. These ranges will be compared to data gathered in the field in a program being carried on at MOTS to evaluate the Scripps visibility model. The summary report will be Issued by NOTS late in 1965.

The comparison cf field test results to simulator results described in Ref. 63 is the only such check on direct visual performance discovered in the literature. Similarly, the NOTS-Scripps studies will produce the only known check on a mathematical model.

A number of studies using data available in the literature have been conducted, examples of which are Ref. 72 and 73. In Ref. 72, the detection probability of a tank seen end-on has been computed using specified meteorological visibility, sky-ground brightness ratio, duration of fixation time per glimpse, target-background contrast, and aircraft velocity. In addition, a confusion factor (CD is introduced into the calculations to account for search when other objects are in the visual field. It is not clear how a value is selected for the factor, however. Curves similar to those shown in Pig. **B are given in the report. Apparently, the model has not been verified by even the simplest experiment.

a« 1 IV I 1 J 1 1 |

0.8 i—

^V CF.I j

04 -

V \ 03 \ N^CF.Z \

\CF'2\CF.. >V \

FT ALT. ^^N^ ^ FT ALT. ^ ^-^

OJ

0.1 ** V UTTKT«

„ mm „ — 300 0J , | l.l, ^-w-'i. 1 > 1 2 » "i A * ? r •

HORIZONTAL RANGE TO TARGET, THOUSANDS OF FT

PIG. »»8. Accumulative Detection Probability Against a JS-III Tank

^ ««^ I in an Area 600 by S00 Yards. The ^""""r*-« assumed confusion factor, CP, is

shown on the curves.

A more complicated model described in Ref. 73 can be applied to recognition of targets by direct vision or with television, infrared, or radar sensors. The complexity is indicated by the use of some 198 symbols

*♦«♦

NAVWEPS REPORT 8617

in this aiodel, including a probability of target recognition by an observer while monitoring some particular sensor display during a particular interval. The formulation of this probability is based upon the work of Boynton et al reported in Ref. U2. The model is programed so that it can be solved on the IBM 7090.

A number of additional reports dealing with visual detection, issued by contractors or government agencies, are listed in the bibliography. Although such reports, including those described above, vary in their content and probable accuracy, they all have one thing in common: their con- clusions or results have not been checked in the field. Furthermore, in some cases the assumptions made are questionable. Choice of experimental data upon which to base model formulation is but one of the problems. As an example, consider some calculations of search efficiency from a moving aircraft that were based upon Ludvigh and Miller's dynamic acuity data. If, instead of Ludvigh and Miller's data, the data from the experiments by Elkin or Berg and Hulbert had been used, the estimated search performance would have been much better.

SUMMARY: SOLUTIONS TO OPERATIONAL PROBLEMS

The necessity of estimating human visual capabilities occurs frequently in weapon feasibility and design studies and in weapon and tactics effectiveness studies. Other programs such as flare development and test also require a knowledge of visual capabilities. In some cases, consideration of geometric or kinematic parameters will suffice. As an example, consider the evaluation of a tactic of delivering a weapon on the first pass over a smalx target from an altitude of 200 feet above the ground. Reference to Fig. 5, 6, and 7 will reveal that the probability of the target's being within view is low until the aircraft is quite close to the target. Unless the flight path is directly over the target, there would probably not be sufficient time to change direction and track the target before weapon release. Hence, it can be estimated without knowledge of human visual capabilities, that the tactic has too high a probability of failure.

Similarly, the frequency and height of cloud cover in typical target areas may preclude the use of certain tactics or systems. If such is the case, to go a step further and assess visual capabilities is unnecessary.

ACQUISITION, DETECTION, RECOGNITION

Most of the time, however, it is not so easy to solve the problem, and a knowledge of visual capabilities is required. The first step is to decide precisely what is needed.

»♦5

NAVWEPS REPORT 8617

Too often terms are used without being even superficially defined, leading to ambiguous results, particularly when the human system (with its associated subjectivity) is being discussed. The process of delivering a weapon includes searching for the target, detecting an object that could be the target, inspecting the object, and deciding whether it is the target. If the object is accepted as a target, recognition is said to have occurred. The next step is to enter the delivery mode and track or lock on the target. Target acquisition might be defined as the entire process of search, detection, inspection, and target recognition. In some applications, the term target acquisition includes the ability to locate the target (on maps, with respect to some terrain features, with respect to the aircraft, etc.). In other applications, target acquisition is defined to include the lock-on phase.

However, precise definition of these words in this report is not appropriate since no particular situation is being discussed, and one definition is not applicable to all existing data. Current concepts include a probabilistic approach which is perhaps most applicable to the operational situation. The most accurate interpretation of data is based upon such an approach. If a tank has been reported in a certain area, and the searching pilot detects a large dark object in the area, he may be willing to enter the attack mode. If tanks have been shooting at airplanes, :he pilot might be willing to fire at maximum range, before he can actually see enough detail in the object to state at some level of confidence that it is a tank. If the enemy has been using decoys, missile launching may be delayed until recognition at a higher confidence level (say 50%) is reached. If friendly forces are in the area, firing may b<* delayed until the pilot is close enough to say with 100% confidence that the object is an enemy tank. Since the same process occurs in any field test, motivational factors must be considered. Motivational missmatch between field tests and combat is probably inevitable, but efforts should be made to minimize it.

USEFULHESC OF DATA

Where a large number of data points are available, and are normally distributed, reporting the mean and standard deviction is sufficient. In practice, especially with field tests, the number of samples is usually small and normality cannot be assured. It is therefore most helpful to report all the data or to show the distribution of the data (histogram). Tests for significance of effects should be included when applicable. but it is also important to present the data itself. To show that some performance at 100 knots is significantly better (at, say, the 0.001 level of significance) than at 200 knots is not enough for a decision. If performance (say, range at target detection) is improved only 10% by slowing down to 100 knots, it may not be worth the disadvantages incurred, whereas a 300% improvement would be.

1*6

MAVWEPS REPORT 8617

MATHEMATICAL MODELS

It appears that the initial reaction to awareness of th<s need for data on target acquisition capabilities is to derive some equations and calculate all of the necessary data. Factors such as human visual acuity and contrast threshold, lighting, atmospheric attenuation, target size, and target background reflectances, are usually included even by the be- ginner. As the theorist studies the problem further, he includes stimu- lus duration, peripheral acuity (detection lobes), and target shape. The finishing touches are added by throwing in internal visual content (contrasts) of the target, nonuniform background, and search techniques with a dash of probability theory. Often a "fudge factor" is applied to ccver fatigue, motivational, and other effects. If intermediate sensors arc used, applicable parameters such as signal-to-noise ratio, lines per i.nch, and contract attenuation, are added to the equations. With some notable exceptions, the results of these calculations are never checked in th; field.

Independent comparison of the results with existing field data ' however meager) shows that errors of 100% are common. Errors of more than 100% are even more common. (It might be noted that many of these etrors are in the favor of whatever missile system is being proposed at the time.)

Efforts to apply analytic methods to the target visibility problem should continue, and support should be given to those proficient in the field, such as Scripps Visibility Laboratory. Such efforts should 'je recognized as being in an exploratory research phase, however, and should be treated as such. Basing missile design on the results of such calculations can easily result in the waste of more money and time than would be consumed by a properly planned flight-test program that would provide more reliable data.

LABORATORY TESTS (SIMULATORS)

Field conditions can be simulated to varying degrees of fidelity in the laboratory and performance f subjects can be measured. Simulation of the target acquisition process can range from use of abstract, rwo- dimensional displays (containing circles, squares, numbers, letters, etc.), to complete three-dimensional terrain models containing trees, roads, grast-, houses, and an adjustable "sun*1. There are a number of such simulators in this country and many reports have been issued giving results obtained with them. These results are almost never checked, even partially, in the field.

Simulator studies are valuable for determining trends to be expected in the field, provide a valuable background and partial understanding of the problem of planning field tests, and can be used to extend field data under the proper conditions. But using simulator data as the only input to analyses of operational problems (which usually require absolute, not relative numbers) is unwise when there is any feasible alternative.

U7

NAVWiiPS REPORT 8617

FIELD TESTS

A field test is, in most cases, also a simulction of a desired combat situation.

Although it is not often done, the physical situation can be duplicated (as on the Coso Test Range, at NOTS) so that one might say this in not a simulation. Differences in mctivation and stress, for example, do exist, and probably make the re- ts noticeably different from those which would be obtained in war.

In mathematical models, relations lips and values of parameters can simply be assumed, and calculations made. In laboratory studies, factors can be closely controlled and a large amount of data can be gathered. In the field tests, however, adequate contr>1 of the variables is often difficult or impossible, and the number of conditions that can be investigated as well as the number of runs at each condition is limited by practical factors. These factors must be kept in mind when planning field tests as well as when judging the usefulnes«; or applicability of field- test results to weapon analyses.

Although the only real solution to an operational problem is a field test, the material presented in this report is helpful in planning, con- ducting, and evaluating field tests, and might be used to some degree to extend the test results.

The information presented in this report car. be placed in two cate- gories: (1) factors affecting air-to-ground dett:tion and recognition ranges and probabilities and (2) past attempts to determine these ranges and probabilities. The first category is simply a presentation of examples of available data from analyses, field measur'eroents, and laboratory studies. The second describes studies of collection and interpretation of test data for various applications and, in some cases, the use of data from the first category for estimating performance levels in operational situations. The second category, which produces results with direct application to military problems, is by nature more controversial.

It is hoped that the results of the operational studies described or referenced in this report will be useful in answering dasign evaluation, or tactical questions. The curves showing obstruction of view and angular rates of the field can be used alone or in combinat'on with the reported field-test results in various types of operational analyses.

«♦8

BLANK PAGE

NAVWEPS REPORT 8617

REFERENCES

1. Naval Air Test Center. Cockpit Visibility, Measurement of, by H. L. Dodson. Patuxent River, Maryland, USNAS, 22 April 1958. (Report No. 1, PTR SI-5002).

2. ffic<s of Naval Operations. Visual Reconnaissance from Aircraft. Washington, D. C, CNO, 1953. (NAVAEF 0O-8OT-U5).

3. U. S. Naval Ordnance Test Station. Empirically Determined Effects of Gross Terrain Features Upon Ground Visibility from Low-Flying Aircraft, by Ronald A. Erickson. China Lake, Calif., NOTS, 13 September 1961. (NAVWEPS Report 7779, NOTS TP 2760).

U. Human Sciences Research, Inc. Information Available from Natural Cues During Final Approach and Landing, by M, Dean Havron. Arlington, Virginia, HSR, March 1962. (HSR-RR-62/3-KK-X, ASTIA No. 285 598).

5. Brow», Robert M. "Visual Sensitivity to Differences in Velocity," PSYCHOL BULLETIN, Vol. 58, No. 2, (March 1961), pp. 89-103.

6. Chief of Naval Operations. U. S. Navy Marine Climatic Atlas of the World, Vol. I, North Atlantic Ocean. Washington, D. C, CNO, 1 Novem- ber 1955. (NAVAER 50-1C-528).

7. . u. s. Navy Marine Climatic Atlas of the World, Vol. TI, North Pacific Ocean. Washington, D. C, CNO, 1 July 1956. (NAVAER 50-1C-529)

8. . u. S. Navy Marine Climatic Atlas of the World, Vol. IV, South Atlantic Ocean. Washington, D. C, CNO, 1 September 1958. (NAVAER 50- 1C-531).

9. . u. S. Navy Marine Climatic Atlas of the World, Vol. V, South Pacific Ocean. Washington, D. C, CNO, 1 November 1959. (NAVAER 50- 1C-532).

10. Scripps Institution of Oceanography. Seasonal Survey of Average Cloudi- ness Conditions Over the Atlantic and Pacific Oceans, by C. R. Fean. San Diego, Calif., VIS LAB, October 1961,(S.I.0. 61-27).

11. Advanced Weapon System Activity. Cloud Cover of the USSR, by Milton Schloss. Applied Science Staff, 17 March 1960.

12. Central Intelligence Agency. National Intelligence Survey, Military Geography (Chapter II). Washington, D. C, CIA, September 19»+9. (Section 23, Weather and Climate).

13. Johns Hopkins University. Hydrographie Office Information on Sea State and Cloud Cover, by J. H. Meal. Silver Springs, Maryland, Applied Physics Laboratories, 28 November 1956. (CLA Internal Memo 593).

U9

NAVWEPS REPORT- 8617

IM. Bureau of Ships. Natural Illumination Charts, by D. R. E. Brown. Washington, D. C, USN, September 1952. (Report No. 37M-1).

15. Middleton, H. E. K. "Vision Through the Atmosphere." Canada, Uni- versity of Toronto Press, 1952.

16. Office of Scientific Research and Development. Visibility Studies and Some Applications in the Field of Camouflage. Washington, D. C, National Defense Research Committee, 19U6. (Summary Technical Re- port Div. 16, Vol. 2).

17. Naval Research Laboratory. Measurements and Estimates of Sky Bright- ness for All Altitudes of the Sun for Various Altitudes of the Ob- server Above the Surface of the Earth, by E. 0. Hulbert. Washington, D. C, NRL, 8 February 1963. (NRL Report U870).

18. University of Michigan. Scintillation and Visual Resolution Over the Ground, by F.R. Bellaire and F. Ryznor. Ann Arbor, Michigan, Willow Run Laboratories, September 1961. (Project Michigan 2900-293-T).

19. Van de Hülst, H. C. "Light Scattering by Small Particles," New York, John Wiley & Sons, 1957.

20. Duntley, S. Q., et al. "Image Transmission by the Troposphere I." J OPT SOC AM, Vol. U7, No. 6, (June 1957), pp. '499-506.

21. Naval Research Laboratory. Atmospheric Transmission in the Visible Region, by J. A. Curcio and K. A. Durbin. Washington, D. C, NRL, 6 October 1959. (NRL Report 59 68).

22. University of Michigan. Scintillation and Visual Resolution Over the Ground, by F. R. Bellaire and F. C. Elder. Ann Arbor, Mich., Willow Run Laboratories, October 1960. (Project Michigan-2900-l?1+-T).

23. U.S. Naval Ordnance Test Station. Summary of Investigations of Heat- wave Effects on Photographic Images, by R. B. Walton. China Lake, Calif. NOTS, October 1962. (NAVWEPS Report 7773, NOTS TP 275»*).

21». Quartermaster Research and Development Center, U. S. Army. Color Regions of the World, by J. V. Chambers and P. C. Dalrymple. Natick, Mass., EPRD, November 1956. (EP-37).

25. National Research Council. Spectral Reflectance Properties of Natural Formations, by E. L. Krinov. Ottowa, Canada, NRC, 1953. (Technical Translation TT-439).

26. Geophysics Research Directorate, U. S. Air Force. Slant Visibility, by R. Penndorf, B. Goldberg, and D. Lufkin. Cambridge, Md., AFCRC, December 1952. (Air Force Survey in Geophysics No. 21, ASTIA No. AD 3276).

50

NAVWEPS REPORT 8617

27. Ashburn, E. V., and R. G. Weldon. "Spectral Diffuse Reflectance of Desert Surfaces." J OPT SOC AM, Vol. UG, No. 8, (August 1956), pp. 583-586.

28. Glenn L. Martin Company. Maximum Ranges for Visual Target Hetection and Recognition, by G. N. Kremonas and H. Sohn. Baltimore, Md., 15 September 1955. (ER-6M53, ASTIA No. AD 116 686), CONFIDENTIAL.

29. U. S. Armed Forces-National Research Council. The Limiting Capa- bilities of Unaided Vision in Aerial Reconnaissance, by S. Q. Duntley. Ann Arbor, Mich., NAS-NRC, January 1953.

30. Blackwell, H. R. "Contrast Thresholds of the Human Eye." J. OPT SOC AM, Vol. 36, No. 11 (November 19M6), pp. 624-6U3.

31. University of Michigan. Detection Thresholds for Point Sources in the Near Periphery, by H. R. Blackwell and A. B. Moldauer. Ann Arbor, Mich., Engineering Research Institute, June 1958. (ERI Project 2U55, USN BuShips Contract NObs-72038, Index 714-100).

32. Scripps Institution of Oceanography. Visual Contrast Thresholds for Large Targets, Part I, the Case of Low Adapting Luminances, by J. H. Taylor. San Diego, Calif., VIS LAB, June 1960. (SIO 60-25).

33. . Visual Contrast Thresholds for Large Targets, Part II: The Case of High Adapting Luminances, by J. H. Taylor. San Diego, Calif., VIS LAB, June 1960, (SIO 60-31).

3U. . Contrast Thresholds as a Function of Retinal Position and Target Size for the Light-Adapted Eye, by J. H. Taylor. San Diego, Calif., VIS LAB, March 1961, (SIO 61-10).

35. Vos, J. J., A. Lazet, and M. A. Bouman. "Visual Contrast Thresholds in Practical Problems." J OPT SOC AM, Vol. U6, No. 12 (December 1956), pp. 1065-1068.

36. Lamar, E. S,t et al. "Size, Shape, and Contrast in Detection of Targets by Daylight Vision. I. Data and Analytical Description." J OPT SOC AM, Vol. 37. No. 7 (July 19'*7), pp. 531-5U5.

37. . "Size, Shape, and Contrast in Detection of Targets by Day- light Vision. II. Frequency of Seeing and the Quantum Theory of Cone Vision." J OPT SOC AM, Vol. 38. No. 9 (September 19U8), pp. 7«4l-755.

38. National Academy of Science-National Research Council. Visual Detec- tion as Influenced by Target Form, by A. B. Kristofferson. Washing- ton, D. C, NAS-NRC, 1957. (NAS-ÜRC Publication 561). pp. 109-127.

51

NAVWEPS REPORT 8617

39. Ohio State University. The Visibility of Non-Uniform Target-Back- ground Complexes: II Further Experiments, by G. A. Bixel and H. Richard Blackwell. Columbus, Ohio, IRV, July 1961. (Technical Re- port 890-2, ASTIA AD No. 297 069).

'♦O. Boynton, R M. and W. R. Bush. "Recognition of Forms Against a Com- plex Background." J OPT SOC AM, Vol. M6, No. 9, pp. 758-76U.

•♦1. National Academy of Science-National Research Council. Recognition of Critical Targets Among Irrelevant Forms, by R. M. Boynton. Wash- ington, D. C, 1957, NAS-NRC. (NAS-NRC Publication 561), pp. 175- 18U.

•♦2. Wright Air Development Center. Laboratory Studies Pertaining to Vis- ual Air Reconnaissance, by R. M. Boynton, C. Elworth, and R, M. Palmer. Dayton, Ohio, WADC, April 1958. (WADC Technical Report 55-30U, Part III, ASTIA No. AD ]M2 27U).

»♦3. Baker, C. A., D. F. Morris, and W. C. Steedman. "Target Recognition on Complex Displays." HUMAN FACTORS, Vol. 2, No. 2 (May 1960), pp. 51-61.

UU. Wright Air Development Center. Partitioning and Saturation of the Perceptual Field and Efficiency of Visual Search, by C. W. Eriksen. Dayton, Ohio. WADC, April 195U. (WADC Technical Report 5U-161, ASTIA No. AD 40730).

U5. National Academy of Science-National Research Council. Problems in the Design of Sensor Output Displays, by S. W. Smith, Washington, D. C, NAS-NRC, 1962. (NRC Publication, Visual Problems of the Armed Forces), pp. 146-157.

46. Smith, S.W. "Visual Search Time and Peripheral Discriminability," J OPT SOC AM, Vol. 51, No. 12, (December 1961), p. 1462.

47. U. S. Naval Ordnance Test Station. Visual Search for Targets: Lab- oratory Experiments, by Ronald A. Erickson. China Lake, Calif., NOTS, October 1964. (NAVWEPS Report 8U06, NOTS TP 3328).

48. National Academy of Science-National Research Council. Visual Search Techniques, edited by A. Morris and E. P. Home. Washington, D. C, NAS-NRC, 1960. (NAS-NRC Publication 712).

49. Ludvigh, E., and J. W. Miller. "Study of Visual Acuity During Ocular Pursuit of Moving Test Objects I. Introduction." J OPT SOC AM, Vol. 48. No. 11 (November 1958). pp. 799-802.

50. Miller. J. W. "Study of Visual Acuity During the Ocular Pursuit of Moving Test Objects II. Effects of Direction of Movement, Relative Movement, and Illumination." J OPT SOC AM, Vol. 48, No. 11 (November 1958), pp. 803-808.

52

NAVWEPS REPORT 8617

51. Tufts University, Institute for Psychological Research. The Effect of Target Ve'ocity, Exposure Time, and Anticipatory Tracking Time on Dynamic Visual Acuitv, by E. H. Elkin. Hedford, Hass., IPR, February 1961. (ASTIA No. AD 256 891).

52. Berg, A., and Hulbert, S., "Dynamic Visual Acuity as Related to Age, Sex, and State Acuity, J APPL PSYCHOL, Vol. U5, No. 2 (1961), pp. 111-116.

53• British Air M.nistry. The Perception of Moving Objects. I., by W. A. Crawford. Famborough, England, Institute of Aviation Medicine, RAF, July 1960. (FPRC 111U).

5*** Boeing Airplane Company. Dynamic Airborne Reconnaissance Display Effectiveness as a Function of Display Scale, by A. R. Yeslin. Seattle, Washington, Aerospace Division, 2 December 1960. (D2-10212).

55. Douglas Aircraft Company. Dynamic Vision I. The Legibility of Equal- ly Spaced Alpha-Numeric Symbols, by S. Lippert. Long Beach, Calif., Systems Research Section, 12 July 1962. (LB 30961, ASTIA No. AD 285 817).

56. . Dynatr.ic Vision II. The Relative Legibility of Alpha-Numeric Symbols, by S. Lippert and D. M. Lee. Long Beach, Calif., Systems Research Section, 17 December 1962. (LB 31157).

57. Scripps Institution of Oceanography. Predicting the Detection Range of a Target in a Moving Field of View, by A. Morris. San Diego, Calif., VIS LAB, December 1959. SIO Report 59-69).

58 • Minneapolis-Honeywell Regulator Company. The Effect of Rate and Di- rection of Display Movement Upon Visual Search, by L. G. Williams and M. S. Borow. Minneapolis, Minn., MPG, 22 November 1962. (MH- MPG-R-RD 6265).

59. National Academy of Science-National Research Council. Natural Ten- dencies in Visual Search of a Complex Display, by Jay M. Enoch. Wash- ington, D. C, NAS-HRC, 1960. (NAS-tfRC Publication 712), pp. 187-192.

60. Wright Air Development Division. Operator Performance in Strike Re- connaissance, by A. C. Williams, Jr., et al. Wright-Patterson Air Force Base, Ohio, August 1960. (WADD Technical Report 60-521).

51« Hughes Aircraft Company. Advanced Tactical Strike System, Vol. II, Part II: Human Factors (U). Culver City, Calif., Hughes Aircraft Company, April 1961. (Final Report P61-02, Contract AF 33(6)OO-»*20«*9), SECRET.

53

MAVWEPS REPORT 8617

62. U.S. Army Personnel Research Office. Human Factors Studies in Image Interpretation: Vertical and Oblique Photos, by J. Zeidner. Washing- ton, D. C, USAPRO, December 1961. (Technical Research Note 120, ASTIA No. AD 1'81 «423).

63. University of Michigan. Model Simulator Studies-Visibility of Mili- tary Targets ar. Related to Illuminant Position, by D. A. Gordon and G. B. Lee. Willow Run Laboratories, Ann Arbor, Mich., March 1959. (Report 2mu-3i4l-T. ASTIA No. AD 213 U09).

6U, »ti.onal Academy of Science-National Research Council. Field and Simulator Studies of Air-to-Ground Visibility Distances, by H. R. Blackwell, J. G. Ohmart, and E. R. Harcum. Washington, D. C, NAS— NRC, 1960. (NAS-NRC Publication 712, Visual Search Techniques), pp. 211-.'30.

65. U.S. Army Aviation Human Research Unit. Research on Human Aerial Observation, Part I, II, and III, by J. A. Whittenburg, et al. Fort Ruckcr, Ala., 'JSCAC, July 1960. (HUMRRO 1-005).

66. . Training Research on Low Altitude Visual Aerial Observation: A Description of Five Field Experiments, by F. H. Thomas. Fort Rucker, Ala., USCAC, July 1962. (Task Observe Research Memo. No. 8).

67. U. S. Army Ordnance Human Engineering Laboratories. Helicopter Arma- ment Program Air-to-Ground Target Detection and Identification, by C. G. Moler. Aberdeen Proving Ground, Md., January 1962. (Technical Memorandum 1-62).

68. U. S. Naval Ordnance Test Station. Air-to-Ground Visual Acquisition of Tactical Targets, by Ronald A. Erickson. China Lake, Calif., N0TS, 3 November 1961. (NAVWEPS Report 780U, N0TS TP 2801), SECRET.

69. Ballistic Research Laboratories. Analysis of Data Collected From an Experiment Involving Low Altitude Reconnaissance and Simulated Acqui- sition of Targets with Rotary Wing Aircraft. Aberdeen Proving Ground, Maryland, April 1962. (Project Thor Technical Report No. »»9, ASTIA No. AD 329 871).

70 * U. S. Naval Ordnance Test Station. Target Detection and Recognition Data Obtained During Flight Tests on the Coso Test Range, by L. 0. Erwin. China Lake, Calif., N0TS, May 1963. (NAVWEPS Report 8127, MOTS TP 3218), CONFIDENTIAL.

71. Scripps Institution of Oceanography. Predictions of Sighting Range Based Upon Measurements of Target and Environmental Properties, by Jacqueline I. Gordon. San Diego, Calif., VIS LAB, September 1963, (SI0 63-23, Contract N123(60i3C) 29657A).

5U

NAV.VEPS RI.PORT 8617

12. General Dynamics, Convair. Visual Detection From Aircraft, by Asb- jorn Linge. San Diego, Calif., Convair, December 1^61, (ERR-SD-150).

73. North American Aviation, Inc.. A Mathematical Model for Predicting Target Identification System Performance, by G. 0. Omstein, R. W. Brainard, and A. B. Bishop. Columbus, Ohio, NAA, 1 February 1961, (NA61H-29).

BIBLIOGRAPHY

MATHEMATICAL MODELS

1. Scripps Institution of Oceanography. A Study of Visibility of Ships at Sea (U), by Jacqueline I. Gordon. San Diego, VIS LAB, June 1961. (SIO 61-16, ASTIA No. AD 33U 57.*L). CONFIDENTIAL.

2. . The Visibility of Nuclear Submarines in the Artie IU), by Jacqueline I. Gordon. San Diego, Calif., VIS LAB, June 1960. (SIO 60-U2, ASTIA No. AD 33M S38L), CONFIDENTIAL.

3. Geophysics Research Directorate. Slant Visibility, by R. Penndorf, B. Goldberg, and D. Lufkin. Cambridge, Md., AFCRC, December 1952. (Air Force Survey in Geophysics No. 21).

U. The Rand Corporation. Target-Search Capability of a Human Observer in High-Speed Flight, by Doris J. Dugas. Santa Monica, Calif., De- cember 1962, (RM-3223-PR).

5. Autonetics. Vision from Low-Flying Aircraft, by Research and De- velopment Division. Anaheim, Calif., 6 April 1962t(EM 1162-103).

6. Royal Aircraft Establishment. Air-to-Ground Visibility for Bomb Aiming, by D. J. Walters. Famborough, England, April 1951. (Tech- nical Note IAP. 997, ASTIA, No. ATI 109 91U).

7. university of Michigan. An Operations Analysis of Aerial Visual Surveillance (U), by W. M. Kinkaid, D. E. Lamphiear, and H. R. Black- well. Ann Arbor, Mich., ERI-VRLf July 1958. (Project Michigan Re- port 2m-282-T, ASTIA No. 11, AD 300 67,j), CONFIDENTIAL.

3. Duntley, S. Q. "Visibility of Distant Objects." J OPT SOC AM, Vol. 38, No. 3 (March 19U8), pp. 237-2U9.

9. Canadian Armament Research and Development Establishment. Visibility Through Television Systems, by D. S, Galbraith. Valcartier, Quebec, CARDE, November 1961. (CARDE Technical Memorandum 663/61).

55

NAVVEPS REPORT 8617

10. Ryll, E., and R. M. Stevens." Visual Aerial Observation for Nap-of- the-Earth Flight Pathr," presented at the Visual Scare.. Symposium, N. E. L., San Diego, Oalif., April 17, 1962. (Operations Research Department, Cornell Aeronautical Laboratory, Inc.).

11. Royal Aircraft Establishment. Air-to-Ground Applications of Visual Detection Lobe Theory, by E. Heap. Farnborough, Engle-nd, January 1962. (Technical Note ARM.715. ASTIA No. ^ 27U 593).

12. Bush, W. R., R. B. Kelly, and V. M. Donahue. "Patterr Recognition and Display Characteristicn, IRE Transactions on Humar Factora in Electrot 'rs." Institute of R?dio Engineers, New York, March 1960.

13. Boynton, R. M., and N. D. Miller. "Visual Performance Under Conditions of Transient Adaptation, Illuminating Engineering." I.luminating En- gineering Society, N»äw York, April 1963.

14. University of Michigan. Effects of Dependencies Among- Elements of Luminance Microstructure Upon Vijuai Form Discrimination, by E. R. Harcura. Ann Arbor, Mich., UMRI, October 1958. (UMRI Project 26U3).

15. Wright Air Development Center. Complexity of Contour in Recognition of Visual Form, by James Deese. Dayton, Ohio, WADC, February 1956. (WADC Technical Report 56-60).

56

i c 0 D <

1 1 1 iX»

J 4 « bO ^ » 94 4) bC --Ni 0) « | U. 1 c > h « M U • C > ^ «>

1 »r1

J • P IB • fJ .H 41 4» -H •H O. 0) 'H 4) V

•> H IB • ♦-» «H 41 41 -H Is* > H

* in w 3 > he

512 iVJqS CO M g > M H IO 4) V B b OH X « -H ß 10

"^ as o ^j to 4-» «H ^ * "^ X O 4-* bO ^ 'H ^ s c x t. « x: g Z (j U X

* UH «0 41 ♦J 4> « i * ^ « « «J « 1 • • OP 'Hi

* a " V • • 0 4-» 'H J3 1 a hp r^ fc > v»

H ^ >, PI 5 g •i in r^ fc > *-•

l! «H S "i •H H ►> t. 0 1 «" H «O C 0 UH U ^i

\ It CO M (^ .HO Ü «1 •H «O C O ^ W «4-1 0

> >H •B 0D M M-i «HO ! ^ ^H U B TD H i

rt * b » U M 0) V

v o v b 3 bo « •0 «t W 4) 4» «J j os d « 5 v u

9 f x: 4) P • o fci W O M 4) 4J

tio

n

s:

An q

c 1 •H A J OS 3 M X P b

! « ♦J « V) M H C T) H 1 c o- >H to m c 9 »•H U -O 3 ü 3 bO x 5 n o c

< L > ID G) j X 0 £ 01

O •0 H •0 IO 1« rH C "O H

CO. -H HJ iB C 3 •H W T) 1 U 3 OO

O 5 P »H f £ S < L > IB M X O X «

• w a. TJ 4-' 10 x;

•""—S. i

CO tN

n H

p 44 CO M 0

• . 4» H «tj • w (XTJ P « ^: H «H

O g .Si^" q *> 1 0 • * 3 «o P

si M • Q O 4) IB bO u 9 09 • O P 4) 10 DC JC aw M & 4-' x c 0 a M 'H bo c M ••-•

l JC aw m ü v x: c O D. M »H hfl C •»J -H

•H U- X! 1 4» C 2 § ID •HI

6 *■' C H •H U. X 1 4> C fctOMH O 01 ^.«H U ^ CO 4-» 4) 4-" e IS p« tr> M H o n ^»H

M «n (O ^ 4) P e to i & .H C.

•p d 0 H

| H^ ^ • fc ftrH 4»

< • J H «H 'H ^J H3 < • J H -H -H ♦J

^ »n ü O « 4) ^ 4>

•O U3 g * C l3 T3 r-i o> 5 K »^ id ^ N *n u O »0 4» X 4>

T» »O X < C 10 T) ^H cr 5 « »^ 10 x « m iB d

as 2 IB H P 0 O U c »w «to 5 ^^ PQ W 4> o.^

SB g rO rH HO 0 U c »co w fc 0 w ^

co ;• (0 >1 O ^^ (O W 4» CU1*-!

OS t m < 4-» > m * o u v "a «a O U V Xi

j 3 ^1 t ro Ö.CJ P B >> p «O 4> 3 X C

3 x t ro ao V «B

1 •. ja «-s U

1 IX ^ 0)

»< 4) bO u « 3 41 bfl (4 41 ^ U. 1 c > II a > u, • c > V 'H

> CU W 'H 4» P • P I« • f» «H 4» >

IO. M .H 4) 4-' • P « • 4J 'H «» ^8- 1 | to w g > u)

H w « P e h of- x: 4» -H c io

w O 4 )|lO U 10 > V) UP 10 4> 4-' ß fci OP X 4J -H C «0

w O

a- * 1 "d 2 O ^ bfl 4-> 'H V X X O v bO 4J «H 4J H x ^ w x *• E X (< 0) X

1 • »4- (0 4) 4-» 4» * 1

« • (p (0 4) V 4) TP 1 • • O 4J 'H JC b 1 • • OP 'H x: <A •♦H r» U * V s «r IUH r- h » P

! H •H H >, fc O (3 •H •H H >, f4 0 H «O C O t4^ M t4H

4 ic ao «3 ip «HO

0 \ i ^ r-l ID C 0 «P M ^ W

^ (j CO IB **H »HO ^H > H 1 -5 U B T) i-i Io e TJ

! 1 f £ 4» P 4) It l! ^J X 4) P 4) \ » U (0 O W 4» P

4> O 4» (4 3 60 «B «IM CO U 10 4) 4-> C

4) O 4) ^ 3 60 10 < i e i JC a n io ^ ^ c ^ a «> iB (4 (4 i 0 I

1 -H • 10 V « 4) V IB

»J PS 3 w XI v ^ 0

•H •• P« 3Sil2ö V M u IB V U IB ^ 10 W W H C TJ H /-" —* IQ ♦J IB IO « ^ C TJ H /^~~N

st

St CO. 'H IB IB C 3

•H U TJ 3 Ü 3 & o P 41 (O bi

1 C O, «H «0 a C 3 J-H u -O 3 U 3 & O x: S w 5 c

a 5 v -w tJ c S < b > IB b()

v_y U P iB « H

1x5 iA 5 c O > t' ,H P f1 "B

< (4 > 10 M -^ « Z O X 4> 4) i X OX 41

1 • w O. X) v> IB X H y • w a-o P "B x H >P CM

V O • R 3 'O P O

« 0 • 2 | T3 P U d ü • Q 0 4> IB bfl

^ aw w £ p x: c v( CL •-• •■-* M C *J -H

O C W ' Q O 4) «0 bO

i s| 13 'r- j«: aw M L. ^ x c Ü CLM'H hoc PM

•H u. x r 8 0 ! S *i •H U X 1 4» C J P y <* iMlOMHOW^-H U] m CO f 4> P B

CO 1 f. «H c

■e u fctOMi- ooj >^'H U m (O P 4> P B

i (O 1 f, -H C < • IU a -H 41

< • J I- -H 'HP »n O O IB 4> X 4)

x> «o x < £ <5 T} H a> 3 ix MH 10 ua

h «d P « O H 6 4J

d < • u an 4i < • »3 H •-< «H fj

io O O 0) 4) XJ 4» -O to X < C IB -O

V H <a H a

i « 1 IB > -d > H «a «g r-t Ol 3 p2 ^H IB ua 53 j % ^ «0 ^ P O O fc

c "to n u o ■0 H HO O U C • rO 0) (4 0 11^^« 0}^ 0.^4

1 M n • -3

1 w >1 o ^^ eo w 4i a»p • «H

CO > 1« ^ U »4 •) T3 io a- 0 u v T}

>% p to a» 3 x: a s Ja b co ao v <o fr ? *> « 3 X C

3 A fc r» 6.0 *-• «

1 ' 1

1 U CO ^ •H 4) «H O UH 0) i • *J r-i (0 c a-H "0

Op

ref

sed

obt

ents

ome

n, o

est

pare

• •woiewoboe u c a £ •* -H c o V O O JUT34J'HU >'HWC4)Cf0C

«•«•OUX 3(0C M 3 41 * ß *» «> ^5 C 4) fO "O -H A 04)f»*-'>,a)woc

t.fj TJO'HM-IMH> •H «Ö P 4-> ^ 0 O -H 10 O .C IQ U M

U «H U b M M M 4)'H4J'H04»4>T>4)

.x:ôx:^TJrH6C NAJ « fl> * id ax: «o /

mosph

1 det

from

of 1

n are

Exam

1 met

ance I

N

o +J <t ^ o m E 4i<03wa>'H «oC a w n J3 v c -H o U •'•HOgU4»*J,4-i

r- 4>t>c94»>fOPi t< X> 0) C 4-» -H £ 4)

3 «4-. ti 4) bO 4) a r-t «D CD x: o o 4) < TJ x:

♦J -H »^ 0» 4J J2

> w « t • <B fl E tn t o fci T) 3 (d a •O W OT34)WWTJ4) 6 4;4->4JCP'H+JC«

as LOO«0'H>H«fl 3 v m u 3 **H u a.»^ >, bo w • o en

a. flj W H » C 4) •-< 4)l0 4)^{;-Ht.TJ«

g B U 4) (0 C C 4) •H C «H S 4) O 4-»

« io m u x> v JC •■< to (0O4-'J34)O4J»JE

< X

1 1 1 1 C <D

1 U CO ^ •H 41 .H Ü UH 4) 1 •

Opt

rcfl

ed

obta

nts

me o

, op

esti

ared

w 4) o c a,

p, c 3 ja -H -H c 5 4>00 P T) -u »H Ü 5'Hwc«»>c<oc t- V'H (0 Q.«0H.HTJ 4>(OT3UX 3«0C W 3 «) » E 4-' 10 Ä C 4) «J TJ «H ja oa)fci4J>.«woc

♦J m <o p ^ t» tt4-» •T30'Ht4-i»4H>

•H « C 4-' Pi O O «H «0 O .C 10 U bO

U 'H Ü U 4 V> V) •J.HÎ.HQ4»4)'04»

/* \v 4*4»»«0i., ax:ie ] jr ♦J H 4» B 4-»

V_ Jo CL v e (* to u v

o COM C 4-' e r-i m c ^H ä Ü fi f0 Ô "0 p 4in>3U4)-H • o C a »»»ÄPC'HO M •«.HüB04>4-'MH

r- t»y>>r34)>«ot< t -ü «) c 4-» «H e 4»

3 UH fc 4) bO 4) a >-* (O CO J= O O 0) < T3 X

4J ^ »H 4» 4-» ^ •HO'-.}» Hbigu

O

fr •OMO'O4)M(0'O4> 4>4^4-'CV'HîCM LuU)0>H>'H<a a: 3 4) 10 U 3 >4H

W) U p.««-) >, bO W • O

SiOW^t.'Ht!<0W CL,

5 6 U 4) (0 C C 4» ^ C -H C 4j O V

M (0 m t. X) « £ -H IB tO U4-'^4)U4->4->e

> < SB

io w C 4)

4» O 4-» X «H Hi 4-' 4-» 6

I I O

•H 4)

a«4H

VISUAL DETECTION OF TARGETS; ANALYSIS AND REVIEN · NAVWEPS REPORT 8617 NOTS TP 36U5 COPY 6 2...

Documents

Transcript of VISUAL DETECTION OF TARGETS; ANALYSIS AND REVIEN · NAVWEPS REPORT 8617 NOTS TP 36U5 COPY 6 2...