Study On Technique Of Opto-electronic Hybrid Joint Transform Correlation Pattern Recognition

The technique of opto-electronic hybrid joint transform correlation pattern recognitionhas been widely applied in several domains, such as automatic control, because of itsadvantages of high speed parallel processing and programmable control. Real-timeperformance and distortion-invariable performance which were deeply analyzed in thepaper were two important indications of the opto-electronic hybrid joint transformcorrelation pattern recognition system.The principle of optical joint transform correlation recognition system based on DSPwas stated in this paper at first. The hardware platform of recognition system was built andthe control and recognition program was written. Then, program was optimized byassembly language in order to improve program efficiency and enhance real-timeperformance of recognition system.In this paper, a multiple-reference-images recognition technology based on the opticalInverse Fourier Transform was presented. For the purpose of improving real-timeperformance further, Inverse Fourier Transform was realized by optical technique insteadof DSP. At the same time, the multiple-reference-images recognition technology canenhance the distortion-invariable performance. An algorithm of evaluating thecomparability of cross-correlation peaks was introduced to distinguish true targets fromfalse ones.The experiment was designed to prove the multiple-reference-images recognitiontechnology based on optical Inverse Fourier Transform and the algorithm of evaluating thecomparability of cross-correlation peaks. The results of the experiment confirmed thefeasibility of the new technology and algorithm.At last, the multiple-reference-images recognition technology based on optical InverseFourier Transform and the algorithm of evaluating comparability of cross-correlation peakswere analyzed. The problems of the technology and algorithm may exist and solvingmethods were pointed out.