Background Clutter Suppression Technology Based On Genetic Neural Network

It is very difficult to detect and track the dim point targets because of the long distance and the strong background clutter noise in the infrared search and track system (IRST).The background clutter is must to be suppressed before targets'detection. So this paper focuses on the estimation and suppression of background clutter of images which involve dim targets.With the powerful nonlinear mapping ability of neural network and global search capability of Genetic Algorithm (GA), a kind of image background clutter suppression technique based on neural network optimized by genetic algorithm is presented in this paper. The parameters of the BP neural network and RBF neural network are optimized by genetic algorithm respectively at first. Then the training samples are sent to the BP neural network and RBF neural network respectively in order to train the neural network. The test sample which is selected randomly is sent to the trained neural network respectively in order to verify the performance of the suppression of background clutter.Gaussianity and independency of residuals are also verified using Kendall rank correlation and Friedman statistic methods after the suppression of background clutter. The performance of the suppression of background clutter is also compared with Uniform weight function. The experiments show this method is feasible and efficient.