Cat-eye target recognition with image registration of dual-CCD

摘要

With the development of optoelectronic surveillance, tracking and optical communication technology, high efficient and reliable target recognition algorithm, especially cat-eye effect target recognition algorithm, becomes more and more important. However, the accuracy and efficiency have been difficult to guarantee. To solve these problems, a cat-eye effect target recognition method based on image registration (CEIR) of dual-CCD is proposed. In this method, before difference operation, image registration is applied to correct corresponding mismatching points. Besides, in the CEIR, an improved algorithm for the exhaustive algorithm is proposed to shorten the running time and improve the correctness of the exhaustive algorithm. Experimental results show that the CEIR is capable of recognizing cat-eye effect targets in static and dynamic background. The CEIR can remove dynamic background distraction and extract cat-eye effect targets with only one pair of images. The CEIR is more efficient in the running time than the shape-frequency dual criterions (SFDC) method. (C) 2015 Elsevier GmbH. All rights reserved.