用全光联合变换相关实现目标识别

Using all-optical joint transform correlator to realize target identification

本文提出了两种改善全光联合变换相关器性能的方法 ,其一是对输入面上的待识别图像和参考图像进行图像分割 ,以得到更高更尖锐的相关峰 ;其二是用多参考图像 ,产生一个标准的相关峰 ,为确定待识别图像和参考图像相关程度提供了判别依据 ,使进行目标识别更加简洁方便。通过实验证明了这些改善方法的可行性 ,并给出了实验结果。

Two methods using to improve the performance of all optical joint transform correlation(AOJTC)are put forward in this paper.The one is to segment images which are the identifying image and the reference images on the input plane,so as to obtain higher and sharper perk of correlation.The other method is to use multiple reference images,which will produce a standard peak of correlation so as to determine whether the identifying image is correlative with the reference image.So target identification is more simple and direct.The feasibility of the two methods is proved through experiment,and obtained the results of the experiment.