| Optical correlation detection is mainly applied optical correlator for target identification. It combines the high speed, parallelism of optical system and the programmability of the computer which has a very important application in the field of optical correlation.However in the practical application. Due to the target image has the distortion (scale and rotation distortion) relatively the reference images, the correlation peak will weaken rapidly, even without correlation peak, leading to the target cannot be accurate identification and positioning. So the key of the optical correlation detection is how to solve the problem of distortion of target recognition. Only solve this problem, the optical correlation recognition technology will have wider application.In this paper, based on optical correlation principle. For further study of the synthetic discriminant function and its derivative algorithm. Finalized with the improved algorithm of maximum average correlation height(MACH)as the core of this paper. Through the optimization of the existing Maximum Average Correlation Height filter, apply it on the Vander Lugt Correlator(VLC), the recognition result is a correlation peak. the correlation peak position determine the target location, the distortion of target recognition was realized. Then, edge exraction the training image and target image by using Sobel edge detection. with the improved maximum average correlation height (MACH) algorithm synthesis of reference images. Apply it on the Joint Transform Correlator(JTC), recognition the distortion of target. The recognition result is a pair of the correlation peak. The distortion target can be accurate recognition.Through the result of computer experiment and the optical experiment show that the improved maximum average correlation height (MACH) distortion has stronger correlation output, has the ideal distortion tolerance, has a strong ability of de-noising at the same time. The distortion target can be accurate recognition. |
PDF Full Text Request