实时联合变换相关图像识别系统的设计  

The Design of the Real-Time Joint Transform Image Correlation Recognition System

作　　者：王红霞 孙红辉 赵选科 王莲芬 张清华 

机构地区：[1]火箭军工程大学,陕西西安

出　　处：《应用物理》2019年第7期348-355,共8页Applied Physics

摘　　要：设计和构建了基于联合变换相关的光电混合图像实时识别系统,该识别系统用电寻址空间光调制器完成数字图像和相干光图像变换,由光学部分高度并行快速完成二维图像的傅里叶变换,平方律探测器CCD完成联合变换功率谱和相关结果的探测。采用VC++.net作为开发语言编制了控制软件,操控空间光调制器和CCD有序工作,实现了目标图像和参考图像的实时相关识别。定义了三个相关参数,可以对相关结果进行定量分析。利用该系统对实景图像进行了识别实验,完成一次识别用时约34 ms左右。The image recognition experimental system of photoelectrichybrid based on the real-time jointtransform correlation is designed and built in the paper.The transform of thedigital images to coherent light images are realized by electro addressedspatial light modulator.The parallel and highspeed processing Fourier transform of two-dimension images is completed by the optical section.The powerspectrum of joint transform and thecorrelation results of the image recognition system are detected by thesquare-law detector CCD.The control software isprogrammed based on VC++.net programming language to control the operations ofthe spatial light modulators and the CCD orderly one by one in order andrealize the real-time correlation recognition of the reference image and thetarget image.Three correlation parameters are defined to do quantitative analysis of the correlationresults.The recognition experiments of real-world scene images based on thesystem are realized and the recognition time is about 34 ms.

关 键 词：实时联合变换相关 图像识别 光电混合 空间光调制器 CCD 

分 类 号：TP39[自动化与计算机技术—计算机应用技术]