Unsupervised Change Detection in Remote Sensing Images Using Pulse Coupled Neural Network

In this paper we propose a context sensitive technique for unsupervised change detection in multitemporal satellite images of same scene using Pulse coupled neural network. PCNN is a novel artificial neural network which is developed in 1990 and based on visual cortex of cats. The binary images generated by each iteration of the PCNN algorithm create specific signatures of the scene which are compared for the generation of the change map .As a case study for the unsupervised change detection multitemporal images acquired by cartosat of Uttarkhand showing recent flooding area in Kedernath brought by heavy rainfall and multitemporal optical images acquired by Landsat on a part of Alaska are considered results are also shown on images acquired on sardina island and on Landsat images of Mexico. Keywords—Pulse-coupled neural networks (PCNNs), unsupervised change detection, multitemporal images and remote sensing.