The System Design And Research On Optical Electrical Hybrid Joint Transform Correlator

Pattern recognition technology is of great importance and has found applications in the fields of milltary target recognition,automation control and finger-print recognition. The optical pattern recognition has the advantages of being with superior high speed and parallel proeessing,which makes it a very important way to realize pattern recognition in real-time.However,its disadvantage is of low precision. On the other hand, the digital pattern recognition has the advantages of being with high Preeision, flexible storage,easy estimation and control, but it has disadvantages,including massive storage requirement and the slow proeessing speed.Based on the combination of these two methods, a feasible optical-electrical hybrid joint transform correlator is proposed in this Paper.In the paper, the overall structure of photoelectric joint transform correlation system is described and the Joint Fourier transform theory is introduced detailedly. the correlation peak position , the orientation of input image and the system parameters are analyzed ,and then the suitable location of the reference objects and the target objects in the spatial light modulator is given.We establish a joint transform correlator systems, optical preliminary tests show that this system can better identify the objects. By using the MATLAB language, we can simulate the correlation system, research the effect of correlation when the image moves, rotates,and changes ,and then make a conclussion that the recongnition results are good when the reference object relatively changes 10% within target and the angle changes less than 4 degrees, and shifting the image does not affect the correlation results.The real-time image's object is ambiguity beeause of complicated background, after studying a variety of image pre-processing technology, in order to improve the recognition, the methods which the input image enhances and sharpens and then enteres a joint transform correlator to deal with are brought forward .Through MATLAB programming, a lot of comparing simulation results show that the method can effectively improve the recognition. In order to solve rotation and different ratio between the target image and reference image, for the disadvantages of current distortion recognition technology, optical interference method is proposed to identify distortion objects .By a large number of simulation experiments , experiments indicate that this method can identify distortion objects effectively.