基于熵加权的多向局部差分和方向梯度的红外小目标检测  

Infrared small target detection based on entropy-weighted multi-directional local difference and directional gradient

作　　者：郭亮 GUO Liang(Security Department,Shanghai University of Engineering Science,Shanghai 201620,China)

机构地区：[1]上海工程技术大学保卫处,上海201620

出　　处：《智能计算机与应用》2024年第6期240-244,F0003,共6页Intelligent Computer and Applications

基　　金：国家自然科学基金(62275153,62005165);上海市产业协同创新项目(HCXBCY-2022-006)。

摘　　要：复杂背景下的小目标检测一直是图像处理领域的热点和难点问题。由于背景复杂、信噪比低等因素,现有方法无法对淹没在强杂波和噪声中的目标进行鲁棒检测。本文提出了基于熵加权的多向局部差分和方向梯度方法。首先,利用熵加权的多向局部差分构造目标显著图,突出小目标,同时抑制背景噪声。然后,通过方向梯度计算,精确计算小目标的边缘信息,定位小目标。最终将目标显著图和方向梯度图融合,得到红外小目标检测结果图。实验结果表明,本文所提出的方法可以抑制杂波并产生更好的结果,并且检测率和误检率都优于对比方法。Small object detection under complex background has always been a hot and difficult problem in the field of image processing.Due to the complex background and low signal-to-noise ratio,the existing methods cannot robustly detect the target submerged in strong clutter and noise.In this paper,multi-directional local difference and directional gradient methods based on entropy weighting are proposed.Firstly,the entropy-weighted multi-directional local difference is used to construct the target saliency map to highlight small targets and suppress background noise.Then,through the direction gradient calculation,the edge information of the small target is accurately calculated and the small target is located.Finally,the target saliency map and the orientation gradient map are fused to obtain the infrared small target detection result map.The experimental results show that the proposed method can suppress clutter and produce better results,and the detection rate and false positive rate are better than the comparison methods.

关 键 词：红外小目标检测 多向局部差分 方向梯度 

分 类 号：TP312[自动化与计算机技术—计算机软件与理论]