提出一种提高联合相关识别的有效方法,该方法通过对识别图像和联合变换功率谱分别进行二次微分,抽取识别图像的特征信息,增强联合变换功率谱中调制条纹的对比度和调制条纹的结构,有效地提高了联合变换相关器的相关输出;在此基础上,构造了基于双二次微分的1f实时光电联合相关识别系统,实验结果表明:该方法较二元联合变换相关,提高了图像相关识别能力。

An effective technique to improved joint transform correlation based on double second order differentials (2DJTC) is proposed. The process is able to extract image feactures and increase the contrast and extent of the fringes structure of joint transform power spectrum. The correlation performance of the proposed 1f real time optoelectronic hybrid correlator based on second differentiation (2DJTC) and compared with the binarized nonlinear joint transfrom correlator (BJTC). The results indicate that 2DJTC produces a reansonably good correlation performance.