- Dr. Habib at the Univ. of Calgary is automatically extracting complex building shapes.
- The process involves a combination of LiDAR and stereo-aerial imagery.
- This sophisticated approach could lead to real gains in exploitation productivity.
The results of an ongoing, innovative approach to automating the extraction of building models are currently being reported in this month’s (May 2010) PE&RS journal. You do have to be a member of ASPRS to access this important research paper.
Briefly, Dr. Habib and others at the University of Calgary have developed a methodology that makes use of both LiDAR and stereo images to more effectively extract complicated building models from the scenes. The process first uses LiDAR to create areas of interest and to develop the initial outlines of the buildings’ rooftops. The boundaries of these areas of interest are then refined using the stereo-aerial imagery. The shapes are further refined using 3D geometric and spectral constraints.
The results shown in the article are very encouraging. It is this kind of sophisticated approach that could lead to significant productivity gains in the automation of feature extraction.
