Digital Height Model of Flanders II

Transcription

1 Digital Height Model of Flanders II Tony Vanderstraete

2 Agenda Introduction Point Classification & Quality Control Applications Open Data Policy Conclusions

3 Introduction

4 Introduction Data acquisition aerial campaign over 3 years ( ) winter period (November - April) Flanders Region, Belgium ( km²)

5 Introduction Data acquisition

6 Introduction Data acquisition LiDAR and imagery taken simultaneously Aerial campaign Imagery LiDAR

7 Introduction Source data specifications DHMF II DHMF I Period LiDAR Point density Geometric accuracy Format Imagery Resolution Geometric accuracy 16 pts/m² XY : 0.10m Z : 0.05m LAZ 0.10m XY: 0.10m 1 pt/ 4m² XY : 0.10m Z : 0.07m ASCII n.a. format Tiff

8 Introduction LiDAR data

9 Introduction Data processing Aerial campaign Imagery LiDAR Postprocessing Geometric processing Radiometric processing Classification

10 Point Classification & Quality Control

11 Point Classification & QC The challenge develop highly cost-efficient processes guarantee constant high quality on a regional scale

12 Point Classification & QC Basic point classification distinction between terrain non-terrain water detailed descriptions focus on automatisation manual intervention only in difficult situations supported by the aerial imagery

13 Point Classification & QC Basic point classification difficult situations temporary features (e.g. building sites) dense vegetation water bodies

14 Point Classification & QC Quality control by FGIA smart sampling strategy based on ISO sample areas of 500x500m / 640km² 4 error classes: not-allowed, critical, large, small NOT-ALLOWED CRITICAL LARGE SMALL

15 Point Classification & QC Quality control by FGIA error decision tree

16 Point Classification & QC Quality control by FGIA smart sampling strategy error decision tree continuous feedback & examples largely manual work (5-10 WD/block)

17 Point Classification & QC Aerial campaign Imagery LiDAR Postprocessing Geometric processing Radiometric processing Classification Quality Control Archive

18 Applications

19 Applications Aerial campaign Imagery LiDAR Postprocessing Geometric processing Radiometric processing Classification Quality Control Archive Dedicated workflow 1

20 Applications Easy-to-use products using automatic vectorization and rasterization techniques e.g. raster DTM & DSM (1m, 5m, 25m, 100m)

21 Applications More elaborate products often require additional, semi-automatic classification procedures mainly detailing the non-terrain class with additional classes building vegetation isolated points

22 Applications More elaborate products

23 Applications LOD 1 buildings +

24 Applications LOD 2 buildings (in research) Test: city of Ghent

25 Applications Solar potential maps (in research)

26 Applications Solar potential maps (in research) DSM 0.5 m Aspect 0.5 m Slope 0.5 m Solar 0.5 m

27 Open Data Policy

28 Open data policy In line with the general policy of the Flemish Government, all source data and products are made available as free open data to the public.

29 Open data policy Distribution channels download webservices online applications & tools

30 Open data policy Data portal:

31 Open data policy Height profile service

32 Aerial campaign Imagery LiDAR Postprocessing Geometric processing Radiometric processing Classification Quality Control Archive Data portal Dedicated workflow 1 Dedicated workflow X Workflow LOD 1 buildings DSM DTM Solar potential map

33 Conclusions

34 Conclusions DHMF II highly valuable data for multiple applications 3 years for 1 update ca. 6 mln ( 340/km²) 20 Tb data Challenges cost-effective, qualitative processing of source data cost-effective updates

35 Agentschap Informatie Vlaanderen Tony Vanderstraete, Toon Petermans, Bart Verbeke, Jo Van Valkenborgh