Efficient building change detection in sparsely populated areas using Coupled Marked Point Processes

In this paper we introduce new model for efficient building extraction and change detection in sparsely populated remotely sensed image pairs. In the two time layers interacting marked point processes recognize the buildings while a priori constraints and low level similarity information are involved in the model. The exploration is accelerated through a non-uniform object birth process, which proposes objects in the built-in regions with higher probability based on global statistical image features.