基于输入面形态学预处理的条纹调制滤波联合变换相关器  

Research on FAF joint transform correlator based on morphological processing of input image

作　　者：王勇[1,2] 朱明[1] 

机构地区：[1]中国科学院长春光学精密机械与物理研究所,吉林长春130033 [2]中国科学院研究生院,北京100039

出　　处：《中国光学与应用光学》2009年第5期460-465,共6页Chinese Optics and Applied Optics Abstracts

基　　金：国家863计划"2005AA778032"

摘　　要：为改善联合变换相关器的相关输出结果,提高图像识别能力,提出了一种将输入面形态学处理与功率谱面条纹调制滤波(FAF)处理相结合的方法。介绍了经典联合变换相关器的基本原理;针对联合变换相关器的不足将输入面形态学预处理和功率谱FAF处理相结合,来改进经典联合变换相关器的相关输出性能,详细分析了上述改进型联合变换相关器的参数选取问题。实验结果显示,改进的联合变换相关器相关峰很尖锐,峰值为9.882×108,比传统的联合变换相关器互相关峰值提高了5.761×108,表明将输入面形态学处理与功率谱面FAF处理相结合效果好于单独使用其中一种方法,改善了传统联合变换相关器的目标识别能力。A image processing method combining a input plane mathematical morphology processing with a power plane Fringe Adjusted Filter（FAF） processing has been proposed to improve the recognition ability of a Joint Transfor Correlator（JTC）. The theory of classical JTC is introduced, and then, aiming at its disadvantage, the combination of the input plane morphology processing and the power spectrum plane FAF processing is used to improve the correlation output performance of JTC. Finally, the parameters of the proposed JTC are analyzed and the optimum parameters are selected. Experimental results show that a very sharpness correlation peak with a peak value of 9. 882×10^8 is obtained by the proposed correlator, which is 5. 761×10^8 higher than that of classical JTC. The proposed method improves the target detecting ability of the JTC, for its performance is better than that of any single method based on JTC.

关 键 词：联合变换相关器 预处理 数学形态学 条纹调制滤波 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]