Location Technology of Correlation Peaks Based on Joint Transform Correlator

Abstract

By comparing the similarity of their contour information, hybrid photoelectric joint transform correlator (JTC) is used to recognize the specific target. An effective method to increase the recognition rate is to extract the contour information of target and template and decrease the noise interference in image background. The principle of joint transform correlator is given in this paper. To increase the recognizing rate, a mask algorithm is presented, with which the coordinate of the correlation peaks can be located precisely. Before the mask algorithm, edge extraction with Robert operator is carried out to enhance the target contour information and reduce the background noise. Amounts of experiments prove the mask algorithm can locate the correlation peaks precisely. As an example, an experimental result is given with a car as the detected target. The coordinate of this car is (362, 438). Experimental results confirm the feasibility and effectiveness of the mask algorithm.