面向无人艇雷达图像的连通区域标记改进算法  

Improved Algorithm of Connected Component Labeling for Unmanned Surface Vehicle Radar Images

作　　者：潘选任 王冬姣[1] 叶家玮[1] PAN Xuan-ren;WANG Dong-jiao;YE Jia-wei(College of Civil and Transportation Engineering,South China University of Technology,Guangzhou 510640,China)

机构地区：[1]华南理工大学土木与交通学院,广东广州510640

出　　处：《水下无人系统学报》2022年第1期78-84,共7页Journal of Unmanned Undersea Systems

基　　金：广东省科技项目(2015B010919006)。

摘　　要：搭载雷达的无人艇在水域中航行时,通过雷达扫描水域附近的环境以识别无人艇周围的障碍物,为了区分雷达图像中的障碍物,通常采用连通区域标记算法对不同障碍物进行标记。由于杂波对雷达成像的干扰,造成雷达图像中像素点增加,使用常见的算法处理图像时效果不佳,为此,提出一种改进算法,在常见算法的基础上增加去除杂波像素点的功能,并标记出图像中的最大障碍物。经过实验分析后可知,相比于改进前的算法,改进后的算法减少了标记的区域数量,处理后的雷达图像更加平滑,便于无人艇对最大障碍物进行预警以及采取合理的避碰措施。When an unmanned surface vehicle(USV)equipped with a boat-borne radar navigates the water,the radar scans the environment near the water,and the radar image is used to identify the obstacles around the USV.To distinguish different obstacles,a connected component labeling algorithm is used to mark different obstacles in the radar image.Owing to the interference from clutter on radar imaging,the pixels in the radar image increase;thus,the results obtained from traditional algorithms to process the image are inadequate.Therefore,an improved algorithm is proposed to improve the function of removing the clutter pixels based on conventional algorithms and mark the biggest obstacle in the image.Compared with the traditional algorithm,the improved algorithm reduced the number of marked areas,and the processed radar images are smoother,which is convenient for the USV to detect the largest obstacle early such that avoidance maneuvers can be performed.

关 键 词：无人艇 雷达图像 障碍物 连通区域标记算法 

分 类 号：U674.91[交通运输工程—船舶及航道工程] U665.22[交通运输工程—船舶与海洋工程]