基于新的加权局部熵的小目标红外图像处理  

Small-target infrared image processing based on novel weighted-local entropy

作　　者：王忠华[1,2] 刘建国[2] 邓鹤[3] Wang Zhonghua Liu Jianguo Deng He(School of Information Engineering, Nanchang Hangkong University, Nanchang 330063, China School of Automation, Huazhong University of Science and Technology, Wuhan 430074,China Wuhan Institute of Physics and Mathematics, Chinese Academy of Science, Wuhan 430071, China)

机构地区：[1]南昌航空大学信息工程学院,江西南昌330063 [2]华中科技大学自动化学院,湖北武汉430074 [3]中国科学院武汉物理与数学研究所,湖北武汉430071

出　　处：《华中科技大学学报（自然科学版）》2017年第8期42-46,共5页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61362036);江西省自然科学基金资助项目(20161BAB202037);江西省教育厅科学技术研究项目(GJJ160696)

摘　　要：为了有效地处理低信噪比复杂背景下的小目标红外图像,提出一种基于新的加权局部图像熵的小目标红外图像处理方法.该方法利用小目标红外图像的内在特点,提出多尺度灰度差异算子和局部图像熵算子,然后通过点积运算获得加权局部图像熵,从而有效地抑制红外图像背景和噪声、增强目标,最终大幅度地提高图像的信噪比.仿真实验结果表明:所提方法能高效地处理复杂背景下小目标红外图像,具有一定的理论和应用价值.For the effective processing small-target infrared images against complex background and low SNR,aprocessing method of small-target infrared image based on the novel weighted-local image entropy was presented in this paper.The multi-scale gray difference operator and the local image entropy operator were proposed by using the inherent characteristics of small-target infrared image.And then the weighted-local entropy was obtained through the dot product operation.In this way,the background and noise were effectively suppressed and then the target was enhanced.As shown in experimental results,small-target infrared images under complex background are efficiently processed by using the proposed algorithm.

关 键 词：图像熵 红外图像 多尺度灰度差异算子 小目标 目标增强 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]