ºÝºÝߣ

ºÝºÝߣShare a Scribd company logo

E Cognition User Summit2009 M Wurm Dlr Structure Types C

?
?
T
Trimble Geospatial Munich

This document summarizes a method for deriving urban structure types from remote sensing data. It involves segmenting and classifying very high resolution imagery and digital surface models to extract buildings and characterize urban landscapes. A multi-step workflow segments the images, optimizes the segments, classifies buildings and land cover, then aggregates the results into homogeneous urban structure types at the block level for applications like urban planning, risk assessment, and analyzing local heating potentials.

1 of 20
Mapping urban diversity Derivation of urban structure types by means of multisensoral remote sensing data Michael Wurm German Aerospace Center (DLR) German Remote Sensing Data Center (DFD) University of W¨¹rzburg, Department for Remote Sensing
Global urbanisation
Urban spatial structure - Urban diversity
Characterisation of urban landscape for ¡­ Physical furnishing of the urban area Mapping/Monitoring the urban landscape Comparability Urban planning Added-value analyses Risk- and vulnerability assessment Assessment of local heating potentials Socioeconomic data Discrimination of homogenous areas on block level ? Urban structure types
Landsat Ikonos/DSM Wickop, 1998 and Banzhaf, 2008
Image Analysis Segmentation Classification Aggregation Added-value Analysis Image Input Workflow Buildings SegOpt LCC BType DSM VHR.opt
Building extraction DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height
?
Image Analysis Segmentation Classification Aggregation Added-value Analysis Image Input Workflow Buildings SegOpt LCC BType DSM VHR.opt
Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Segmentation-optimization Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC
Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Classification parameter: Spectral Attributes Shape and area Features Neighborhood Texture Hierarchical Features
?
?
Image Analysis Segmentation Classification Aggregation Added-value Analysis Image Input Workflow Buildings SegOpt LCC BType DSM VHR.opt
Landcover Vegetation density [%] Building density Impervious surface [%]
Schrebergarten Einzel-/Doppelhausbebauung Freistehende Geschossbauten Reihenhaus Zeilenbebauung Offene Blockrandbebauung Geschlossene Blockrandbebauung Offene Blockbebauung Geschlossene Blockbebauung Geschossbaukomplexe Hallenbebauung Punkt-/Scheibenhochhaus Mischbebauung Shape parameters for classification Number of courtyards Sealed areas Vegetated areas Building elevation Building levels Compactness Volume Area Lenght/Width Length of main line Urban Structure Types
Image Analysis Segmentation Classification Aggregation Added-value Analysis Image Input Workflow Buildings SegOpt LCC BType DSM VHR.opt
Case Studies Integration of UST for social sciences Neighborhood Comparison Individual Household
Case studies Local level: Supportive analyses for district/local heating Annual heat demand / development costs Potential for district heating modeled via urban structure Buildings 0.0 ¨C 2.5 [achievable kWh/a per invested €] 2.5 ¨C 5.0 5.0 ¨C 10.0 10.0 ¨C 20.0 > 20
Summary VHR and DSM Detailed classification of urban spatial structure Urban structure types Homogenous areas Added-value analyses Local heating potentials Socioeconomic analysis [email_address]

More Related Content

E Cognition User Summit2009 M Wurm Dlr Structure Types C

  • 1. Mapping urban diversity Derivation of urban structure types by means of multisensoral remote sensing data Michael Wurm German Aerospace Center (DLR) German Remote Sensing Data Center (DFD) University of W¨¹rzburg, Department for Remote Sensing
  • 3. Urban spatial structure - Urban diversity
  • 4. Characterisation of urban landscape for ¡­ Physical furnishing of the urban area Mapping/Monitoring the urban landscape Comparability Urban planning Added-value analyses Risk- and vulnerability assessment Assessment of local heating potentials Socioeconomic data Discrimination of homogenous areas on block level ? Urban structure types
  • 5. Landsat Ikonos/DSM Wickop, 1998 and Banzhaf, 2008
  • 6. Image Analysis Segmentation Classification Aggregation Added-value Analysis Image Input Workflow Buildings SegOpt LCC BType DSM VHR.opt
  • 7. Building extraction DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height DSM VHR Fused Blocks-Import Segmentation 1 Segmentation 2 Building Classification Height
  • 8. ?
  • 9. Image Analysis Segmentation Classification Aggregation Added-value Analysis Image Input Workflow Buildings SegOpt LCC BType DSM VHR.opt
  • 10. Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Segmentation-optimization Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC
  • 11. Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Buildings-VHR Input Basic Segmentation Optimzation Level 1 Optimization Level 2 Optimzation Level 3 Classification Building Classification Streets Import LCC .1 LCC .2 LCC .3 LCC .4 Final LCC Classification parameter: Spectral Attributes Shape and area Features Neighborhood Texture Hierarchical Features
  • 12. ?
  • 13. ?
  • 14. Image Analysis Segmentation Classification Aggregation Added-value Analysis Image Input Workflow Buildings SegOpt LCC BType DSM VHR.opt
  • 15. Landcover Vegetation density [%] Building density Impervious surface [%]
  • 16. Schrebergarten Einzel-/Doppelhausbebauung Freistehende Geschossbauten Reihenhaus Zeilenbebauung Offene Blockrandbebauung Geschlossene Blockrandbebauung Offene Blockbebauung Geschlossene Blockbebauung Geschossbaukomplexe Hallenbebauung Punkt-/Scheibenhochhaus Mischbebauung Shape parameters for classification Number of courtyards Sealed areas Vegetated areas Building elevation Building levels Compactness Volume Area Lenght/Width Length of main line Urban Structure Types
  • 17. Image Analysis Segmentation Classification Aggregation Added-value Analysis Image Input Workflow Buildings SegOpt LCC BType DSM VHR.opt
  • 18. Case Studies Integration of UST for social sciences Neighborhood Comparison Individual Household
  • 19. Case studies Local level: Supportive analyses for district/local heating Annual heat demand / development costs Potential for district heating modeled via urban structure Buildings 0.0 ¨C 2.5 [achievable kWh/a per invested €] 2.5 ¨C 5.0 5.0 ¨C 10.0 10.0 ¨C 20.0 > 20
  • 20. Summary VHR and DSM Detailed classification of urban spatial structure Urban structure types Homogenous areas Added-value analyses Local heating potentials Socioeconomic analysis [email_address]