Research On Target Match Of Missile Terminal Guidance

Missile terminal guiding based on image refers to use some special target matchalgorithm to perform the task of precision strike in the battle with the image captured from theCCD. And the capability and efficiency of the target recognition algorithm is one of the keytechnologies in the research of terminal-guiding. Therefore the paper has high value in theoryand engineering application research.This paper proposed image recognition algorithm based on RBF (Radius Based Function)neural network after the classification of current target recognition algorithms.(1) Classification of currently most used target recognition algorithms is achieved, namelytransformation domain, feature match, gray correlation, model. Analysis of math-modelconstruction, target pixels preprocessing and the performance of algorithm areaccomplished to gain the features and weaknesses of each algorithm.(2) Proposed target recognition algorithm based on multi-resolution and multi-Gabor filterand RBF neural network with the analysis result of other algorithms. The input imagewould be processed by multi-angle and multi-wavelength of Gabor filter. Then thecalculation of11-dimensional feature vector would be implemented and acted as trainingdata as hidden layer and output layer of RBF.(3) The algorithm has been tested on the platform of Cyclone II FPGA and Matlab, includingfloat calculation model, function blocks decomposition, realization of Verilog hardwaredescription language, parallelism and performance analysis. The test result of FPGA andMatlab proved that the algorithm proposed in this paper has good real-timing andeffectiveness.