The Research Of Image Processing In Object Detection And Recognition By Using Joint Transform Correlator

Optical pattern recognition has the advantages of superior high speed and parallel processing, which make it be a very important way to realize pattern recognition practically and in real time. It has been widely studied and used in the fields of military target recognition, automation control and fingerprint recognition. With the high-speed development of modern technology, we have higher requirement in pattern recognition. Photoelectric joint transform correlation system not only have the ability to provide large volumes of information and two-dimensional parallel processing, but also have the advantage of high precision, excellent flexibility, easy to control and judge like digital computers. So photoelectric joint transform correlator has become one of most active research areas in pattern recognition.This paper is broadly divided into two major parts. In the first part of theoretical analysis, we give the theoretical analysis of description of the optical correlation recognition and detection; and then we introduce the joint transform technologies and the system we used in the paper; finally, we introduce detailly the Fourier transform and wavelet transform theory. In the second part, We focused on the image processing of different segments As the distortion of the image in the process of taking and transmissing, we do a series of pre-procesing to the input image, such as gray-scale correction, histogram equalization, wavelet denoising and so on, the input image after pre-processing can get better joint power spectrum. Since the direct current component (grade-zero items) occupies the most part of energy in the relevant surface, and the inter-related energy is relatively weak, we carry out a series of non-linear processing to the joint power spectrum in the paper, such as Laplacian edge detection and binarization. From the effects of image, the power spectrum after binarization can gain better correlation peak images. The existence of direct current item and the cross facula in the correlation peak can influence finding the location of inter-related spot, we adopt the image subtraction, direct elimination and region growing to deal with the influence items. At the same time, we denoise the correlation peak, from the results of the denoising, the method of wavelet de-noising is better than the traditional methods.