In this thesis an automatic approach for vehicle detection from optical satellite imagery is presented. In contrast to stationary sensors like induction loops or bridge mounted cameras, spaceborne imaging systems provide large area views of traffic situations. The proposed approach distinguishes between isolated and grouped vehicles. Isolated vehicles are detected by pattern recognition methods. Hypotheses for vehicle queues are found using differential geometric line extraction. Utilizing robust parameter estimation, single vehicles are determined within those vehicle queues. Additionally, movement of isolated vehicles are estimated. The approach is evaluated on QuickBird images of an urban area. While the completeness is about 65%, the developed approach reaches a high reliability of 95% on average.
«
In this thesis an automatic approach for vehicle detection from optical satellite imagery is presented. In contrast to stationary sensors like induction loops or bridge mounted cameras, spaceborne imaging systems provide large area views of traffic situations. The proposed approach distinguishes between isolated and grouped vehicles. Isolated vehicles are detected by pattern recognition methods. Hypotheses for vehicle queues are found using differential geometric line extraction. Utilizing robus...
»