The Research of the ATR System Based on Infrared Images and L-M BP Neural Network

With the broad application of information processing technology in the surveillance equipment, the automatic target recognition (ATR) technology has become a key part of the battlefield intelligence processing system. In this paper, we presented an approach for building an ATR system with improved artificial neural network, which can be used to recognize and classify the infrared targets in army field. Because of the invariance of rotation, translation and scaling, we selected the features of Hu invariant moments and roundness as input of the neural network. In order to increase the speed of training, the L-M (Levenberg-Marquardt) algorithm was introduced to improve the traditional BP neural network. The results of simulation show that the approach can meet the requirement of the ATR system in high adaptability and good identification effect.