Guidelines for handling large amount of KML data

21
FIRST - Jövő Internet kutatások az elmélettől az alkalmazásig TÁMOP-4.2.2.C-11/1/KONV-2012-0001 Guidelines for handling large amount of KML data Marianna Zichar FIRST – Future Internet Research, Services and Technology TÁMOP-4.2.2.C-11/1/KONV-2012-0001

description

The practice of representing data with geospatial information can be observed in many disciplines and in our everyday life too. Use of free solutions is very popular and usually making the first steps easy. However, what happens if the amount of data exceeds a critical value, and we will face bad visual experiences or the performance of the system dramatically decrease because of the large number of features? Advanced IT skills can help to overcome this issue, but without them, browsing the reference guides and textbooks will not probably help, and we will give up. Presented by Marianna Zichar, assistant professor, University of Debrecen

Transcript of Guidelines for handling large amount of KML data

Page 1: Guidelines for handling large amount of KML data

FIRST - Jövő Internet kutatások

az elmélettől az alkalmazásig

TÁMOP-4.2.2.C-11/1/KONV-2012-0001

Guidelines for handling large

amount of KML dataMarianna ZicharFIRST – Future Internet

Research, Services and Technology

TÁMOP-4.2.2.C-11/1/KONV-2012-0001

Page 2: Guidelines for handling large amount of KML data

• Keyhole Markup Language• Range of potential users• NetworkLink• Region• NetworkLink & Region• Existing solutions• My classification• Conclusions

Overview

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 3: Guidelines for handling large amount of KML data

• XML based language to describe features with geospatial information

• Google Earth / Map • Generable by the application, but

editable• Simple text file with kml (or kmz)

extension• OpenGIS KML 2.2 Encoding

Standard

Keyhole Markup Language

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 4: Guidelines for handling large amount of KML data

• Abstract elements with derived elements

• Inheritance from ancestors• gx: Google extension

namespace

Class tree for KML elements

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 5: Guidelines for handling large amount of KML data

Geometry

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 6: Guidelines for handling large amount of KML data

• Geographers• Surveyors• Medical researchers, physicians• Epidemiologists• Biologists• Linguists• Agronomists• Architects• Catastrophe managers• Retailers • Etc.

Range of potential users – motivation

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 7: Guidelines for handling large amount of KML data

• Use the GUI of GE and save your features• Limited customization• Not all the KML elements are available

The most simple way to create

Page 8: Guidelines for handling large amount of KML data

• Hundreds of placemarks• Large files• Decreasing performance• Editing is hard

• Solution• Divide and conquer• Network link & Region

Issues

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 9: Guidelines for handling large amount of KML data

• Simple element referencing another KML file or KMZ archive

• The fetched KML file can contain• A hierarchy of Feature elements• One (optional) <NetworkLinkControl> element that controls

the behaviour of the source (fetching) KML file.• A hierarchy of network linked KML files can be created.

NetworkLink

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 10: Guidelines for handling large amount of KML data

• For complex sets of KML files, you can partition the KML into multiple files that link to each other.

• When data is changing frequently, the interval refresh feature of <NetworkLink> provides for timely refresh of the displayed data.

• When data is distributed over a large area of the Earth’s surface, view-based refresh feature enables GE to perform a spatial search based on the current viewpoint and load only the data that is currently within the user’s view.

• For very large data sets, the region feature allows you to control which parts of the data set are shown in a given view.

Why to use Network Links?

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 11: Guidelines for handling large amount of KML data

• Refers to a geographical area of 2D or 3D which extension is specified by its bounding 2D or 3D box.

• Can be assigned to any KML feature

• The features of a region are visible if the region itself is visible on the screen and it can be contained by a box with an edge length falling between two values specified as minLodPixels and maxLodPixels.

Region

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 12: Guidelines for handling large amount of KML data

• The networklink element has a viewRefreshMode subelement, which value has to be set onRegion to fetch the referenced file content according to the active status of the region.

NetworkLink with Region

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 13: Guidelines for handling large amount of KML data

• No tool in GE interface to define regions • Regionator• Free Python library to divide thousands of

placemarks, polygons and paths into separate regions for better viewing.

• Requires programming skills.

Existing solutions

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 14: Guidelines for handling large amount of KML data

• KML2KML• Commercial application with a number of tools

for reorganizing and optimizing KML files.• Creates one single region to each checked

feature types based on their minimal bounding boxes

• No possibility to divide the features of the same type to several regions with different LOD values.

Existing solutions

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 15: Guidelines for handling large amount of KML data

• Further commercial applications (e.g. Arc2Earth) do exist.

• Some companies offer services of creating data in KML format as well.

• In spite of these opportunities, my personal experiences show that a practical guide is needed.

Existing solutions

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 16: Guidelines for handling large amount of KML data

The objectives of the most frequent usage methods related to placemarks are:

• To reduce the number of placemarks packed closely together.

• To separate different types of data into layers.

• To fade features in and out of view.

Objectives

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 17: Guidelines for handling large amount of KML data

Regions with the same bounding box are used but with different LOD settings and the goal is to increase the number of visible features from the same area while zooming in.

Features have to be grouped into several folders assigning the same bounding box to each of them.

Case 1.

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 18: Guidelines for handling large amount of KML data

• Increase the performance of rendering when the number of the futures is critical, but the aggregation level is acceptable .

• Features have to be divided into separate files.• The main KML file contains references to the files

with the help of network links. Every network link has an associated region, where the LOD can be even the same, but the bounding box match the physical extent of the features.

• Computer does not have to manage data in memory, which are actually not visible, then the performance of our system will definitely grow.

Case 2.

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 19: Guidelines for handling large amount of KML data

• If the spatial distribution of the placemarks is approximately consistent then a classical quadtree structure is perfect to represent the nested regions.

Case 3.

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 20: Guidelines for handling large amount of KML data

I provided some practical guidelines by differentiating three cases, where instructions and explanations help the users without strong IT skills to manage large amount of data.

Conclusions

5th GIS Conference 29 May – 31 May 2014, Debrecen

Page 21: Guidelines for handling large amount of KML data

Thank you for your attention.

5th GIS Conference 29 May – 31 May 2014, Debrecen