The Design Of Optical Electrical Hybrid Joint Transform Correlator And Correlation Peak Image Process

It need to make matched filter when using the traditional matched filtering method of correlation recognition to recognition image , it is very troublesome, very time-consuming, and cannot achieve the real-time detection of target and recognition by this kind of method. So someone proposes using optical electrical joint transform correlator identify method to compensate for the lack of correlation recognition matched filtering. Optical electrical joint transform recognition correlator combines optical pattern recognition technology information processing capacity, parallel processing, quick speed and high precision, flexible computer programmable sexual advantage, realize the real-time fast recognition goal.The basic theory of optical electrical joint transform correlation system is detailed introduced in the paper, and the traditional matched filter and joint transform correlation respective characteristics is analyzed. Optical electrical joint transform correlator experiment system was built, the experiment device set-up was completed, and the comparison experiments were performed to verify recognition effect of the system. Simulation experiment for image translation, angle change and size change targets. Through MATLAB language, theoretical analysis of the image correlation recognition has no influence on translation, and joint transform correlator has good recognition effect conclusion when reference image and target image the angle in 4 degrees and the size changes in less than 10%. Optical interference method is used in order to solve the reference image and target image between angle changes and size changes, experiments show that the proposed method can be a very good recognition. By this way, the computer's computation has reduced, the time of relevant recognition has reduced, the efficiency of recognition has improved. In order to get better correlation results, correlation peak image information was analyzed. Zero flare and liquid crystal diffraction flare were removed through image minus method and directly elimination method, and other image noises were wiped off with two-dimensions median filter method. The results shown: the quality of correlation peak image was improved by these methods.