Pulse Coupled Neural Networks for detecting urban areas changes at very high resolutions

The development of fully automatic change detection procedures for very high resolution images is not a trivial task as several issues have to be considered. The crucial ones include possible different viewing angles, mis-registrations, shadow and other seasonal and meteorological effects which add up and combine to reduce the attainable accuracy in the change detection results. However this challenge has to be faced to fully exploit the big potential offered by the ever-increasing amount of information made available by ongoing and future satellite missions. In this paper a novel approach based Pulse-Coupled Neural Networks (PCNNs) for image change detection is presented. PCNNs are based on the implementation of the mechanisms underlying the visual cortex of small mammals and with respect to more traditional neural networks architectures own interesting advantages. In particular, they are unsupervised and context sensitive. The performance of the algorithm has been evaluated on very high resolution QuickBird and WorldView-1 images. Qualitative and more quantitative reuslts are discussed.