Applicability of Modified Counter-Propagation for Satellite Image Classification

Abstract

A supervised classifier for satellite images by the Modified Counter-Propagation(MCP) is proposed. The MCP is a neural network for the competitive learning and it is the modified version of the Counter-Propagation, whose competitive layer is replaced by the Self-Organizing map(SOM). The Landsat image data are adopted as the input data of the MCP, and the output layer consists of the pixel values, which represent categories to be classified. Our result shows that the MCP can classify the data more accurately, objectively, and stably than the SOM only.