Introduction
High expectations of mobile subscribers, increasing RF network density, increased capability of RF planning tools to handle high resolution data and increasing temporal and spatial (e.g., 50 cm) resolution satellite imagery availability creates the high demand for 3D Urban Geodata in Telecommunication industry for Radio Network Planning and Optimisation.
Building heights for 3D city model extracted from aerial photo, high resolution satellite imagery and combined LiDAR and aerial image data are popular. However, building height extraction utilising aerial images have generally suffered due to low temporal resolution and the high cost of image acquisition. The arrival of LiDAR data has opened a new level for building height extraction. Although the results obtained using LiDAR data are promising, it is still expensive and needs significant editing and computing power. This makes us (Empower Consultancy Pvt. Ltd; South India based leading geospatial company) to generate 3D buildings using stereo satellite imagery as an cost effective alternative.
Study area
The WorldView High Resolution Stereo Satellite Imagery used was of 2 Km2ofCarnaxide; a town near Lisbon in Portugal covering the co-ordinates between X Max: 38o43’40.2” X Min: 38o42’56.51” and Y Max: 9o14’22.06” Y Min: 9o13’13.40”. The area was chosen for its organised building layouts, undulating terrain, clear view of road networks etc.
Figure 1:Study Area in Ortho Photo
Methodology
A stereo pair of World View 1 imagery was used to build a digital surface model. The difference between this model and the digital terrain model of the study area highlighted the elevations of man-made features. This information was then separated to generate the building model with building vector data.
Figure 2: Flowchart of 3D City Model Generation
The Rational Polynomial Coefficient (RPC) data was used to process the stereo images. A digital surface model of the study area was created through area-based matching technique, which determines the correspondence between two image areas according to the similarity of their gray level values using a 3 × 3 square correlation window. The technique was applied in the Leica Photogrammetry Suite (LPS) software. Figure 3 shows part of the digital surface model created.
Figure 3:Digital Surface Model of the Study Area
Figure 4:3D City Model Generated from Stereo Imagery over Ortho Photo
The Digital Surface Model (DSM) created was compared to the Digital Terrain Model (DTM) representing the ground elevations (excluding above ground features) of the study area. Difference between the DSM and DTM elevation values were extracted and incorporated with building vector model in building height column. Final 3D City Model was generated; Figure 4 shows the 3D City model over Ortho Photo. High priority was given to the building height assessment to exclude the incorrect elevation values produced from the automatic DSM creation step due to image matching and interpolation errors.
Limitations
Automatic generation of DSM creates incorrect elevation values randomly, which needs to be corrected manually by taking the final file again to the stereo model. Selection of suitable stereo satellite imagery makes difference in quality. The GCP points should be utilised to get absolute xyz in output.
Advantages
This method provides a cost effective alternative to the use of aerial photography and LiDAR as processing time is shorter since there is no need to perform complex algorithms or computations. With this method 3 -2 M spatial and vertical accuracy can be achieved with GCP points. This method is suitable for all the types of urban morphology.
Conclusion
We examined the use of World View stereo imagery to extract 3D building information in regards to height and Planimetric position. Digital surface and terrain models were used to extract building heights. The results were compared to building heights available in open source database. Our accuracy assessment results showed that 3-2 M spatial and building height accuracy can be achieved in this method. The results suggest the feasibility of using the World View Stereo models for applications that range from building height estimation for RF planning & optimisation in different urban morphology.
