基于边缘提取的局部对比红外弱小目标检测改进算法  

Infrared Small Target Detection Based on Local Contrast Method Improved by Edge

作　　者：王帅[1] 程洪玮[2] 莫邵文 曾剑[1] 江丹[1] 

机构地区：[1]国防科技大学电子科学与工程学院,湖南长沙410073 [2]北京跟踪与通信技术研究所,北京100094

出　　处：《数字技术与应用》2016年第8期146-148,共3页Digital Technology & Application

摘　　要：针对复杂背景下红外弱小目标检测中存在虚警率较高等问题,提出一种基于局部对比度计算显著图结合边缘提取消除虚警的目标检测算法。该算法首先利用LCM(局部对比度)算法提取视觉显著区域,再利用Canny算子提取云层边缘,将得到的视觉显著图与云层边缘进行对比,剔除掉虚警目标,然后对得到的图像进行门限分割,从而分离出目标。实验表明该算法对多云背景下红外弱小目标具有较好的检测效果,能够解决局部对比度算法存在的虚警率较高的问题。An algorithm of small and dim infrared target detection base on LCM （Local Contrast Method） combing with edge detection to solve the problem of high false alarm in small and dim infrared target detection under complex background. This algorithm firstly use the LCM （Local Contrast Method） to extract the visual salient region in the image. Then the Canny operator is used to extract the edge of the doud. The next we will get the local contrast map to compare with the result of Canny operator algorithm, the target which is located of the cloud edge will be eliminated. The last, adopt to threshold segmentation to separate the target. Analysis of the result shows that proposed algorithm which is based on LCM and Canny operator has a better detection result in small and dim infrared target detection under cloudy background. What＇s more, it can solve the problem of the high false alarm rate by LCM algorithm.

关 键 词：弱小目标 视觉显著性 局部对比度 CANNY算子 

分 类 号：TN976[电子电信—信号与信息处理]