The design of Top-Hat morphological filter and application to infrared target detection

Automatic detection and track for infrared target is of great significance in modern world. In this paper, two novel methods which can develop optimizing Top-Hat morphological filtering parameters are presented for spot target detection. One is based on neural network. Its structuring element is a two-layer feed-forward network which is trained by a mass of sample nets. It regards Top-Hat operation as a whole and one layer, and defines the node of the output layer as the maximum gray-scale image vector after Top-Hat operation. The other is based on genetic algorithm. It adopts the interval discretization code and new primary and secondary mood crossover and mutation. Experimental results show that the identified probability of images (SNR is about 2) can reach more than 98% by this method.