基于双树复小波变换和混沌粒子群优化的红外小目标检测  

Infrared small target detection based on dual-tree complex wavelet transform and chaotic particle swarm optimization

作　　者：吴一全[1] 纪守新[1] 占必超[1] 

机构地区：[1]南京航空航天大学信息科学与技术学院,南京210016

出　　处：《高技术通讯》2010年第9期917-923,共7页Chinese High Technology Letters

基　　金：国家自然科学基金(60872065)资助项目

摘　　要：针对存在背景干扰和噪声情况下的红外图像弱小目标检测问题,提出了基于双树复小波变换和混沌粒子群优化的检测方法。该方法一方面先基于双树复小波变换对原始图像进行去噪,再利用Top-hat算子抑制背景;另一方面先利用Top-hat算子抑制原始图像的背景,经双树复小波去噪后,再进一步使用Top-hat算子。将上述两方面得到的图像求和即为预处理图像。然后基于混沌粒子群优化的类内绝对差及背景与目标面积差的阈值选取方法分割预处理图像。大量实验结果表明,与基于小波和形态学的红外目标检测方法相比,该方法抗噪性强,具有更为优越的检测性能。Aiming at the problem of detecting the dim targets in infrared images that contain the background interference and noises, the paper proposes a new detecting method based on dual-tree complex wavelet transform and chaotic particle swarm optimization. The method is described as below. It denoises the original image based on the dual-tree complex wavelet transform, and then the background of the denoised image is suppressed using the Top-hat operator. At the same time, the background of the original image is suppressed using the Top-hat operator, and then the Top-hat operator is used once more via dual-tree complex wavelet denoising. Adding up the above-mentioned two resultant images gives a preprocessed image. The preprocessed image is segnented using a algorithm for selection of the thresholds of the withinclass absolute difference and the area difference between background and target based on the chaotic particle swarm optimization. Lots of experimental results showed that, compared with the infrared target detection method based on wavelet and morphology, the suggested method was stronger in anti-noise performance and more superior in detection of infrared dim targets.

关 键 词：红外弱小目标检测 双树复小波变换 TOP-HAT算子 混沌粒子群优化 类内绝 对差 

分 类 号：TP274.52[自动化与计算机技术—检测技术与自动化装置]