RS Image PCNN Automatical Segmentation Based on Information Entropy

Pulse Coupled Neura Networks has the essential differences with the traditional artificial neural network in simulating biological visual, so PCNN is widely used in image processing fields. In PCNN model, In image processing, we often use the information entropy as tools to evaluate the effect of image processing, namely the greater the value of information entropy the better the image. The cycle number under the given parameters influences directly the segmentation result. Determining the loop-interaction cycle number at the best segmentation times is a difficult problem. This paper puts forward a PCNN image segmentation algorithm based on the maximum entropy principle. The algorithm determines the cycle number with the maximum entropy in order to realizing the best image segmentation automatically based on regions.