基于加权LoG算子的红外弱小目标检测方法研究  

Infrared small target detection using weighting LoG operator

作　　者：周苑[1] 张健民[1] 林晓[2] 

机构地区：[1]河南工程学院计算机学院,河南郑州451191 [2]上海交通大学电子信息与电气工程学院,上海200240

出　　处：《应用光学》2017年第1期114-119,共6页Journal of Applied Optics

基　　金：国家自然科学基金(U1304616)

摘　　要：提出一种基于加权LoG算子及形态学运算的方法来提高远距离红外弱目标检测效率。通过不同尺度的加权LoG算子对图像进行运算,提取响应值最大的特征图。对图像进行形态学运算去噪,并进行Otsu二值分割聚类。输出目标点在图像上的位置坐标。实验结果表明:该方法与传统滤波方法相比,信噪比增益为36.9,杂波抑制因子为4.7,均比传统滤波方法要好。A new algorithm based on weighting LoG operator and morphological processing was proposed to solve the remote infrared small target detection problem. Firstly, the image was processed by using weighting LoG operators with different scales, and the max characteristic image of response value was extracted. Then, the morphological process was conducted to remove the image noise, and the Otsu binarization segmentation clustering was done. Finally, the position coordinates on the image of target point were output. Experimental result shows that compared with classical filtering method, the proposed method has obvious advantages, the signal to clutter ratio （SCR） gain is 36.9, the background suppress factor （BSF） is 4.7.

关 键 词：加权LoG算子 形态学运算 弱小目标检测 

分 类 号：TN219[电子电信—物理电子学] TP317.4[自动化与计算机技术—计算机软件与理论]