应用激光, 2019, 39 (4): 671, 网络出版: 2019-10-12
基于三维激光扫描的水面目标物体检测
Water Surface Target Detection Based on 3D Laser Scanning
无人船 目标检测 三维激光 最近邻聚类 AABB包围盒 unmanned ship target detection 3D laser nearest neighbor clustering AABB bounding box
摘要
水面目标物体检测是无人船识别水面物体或避开障碍的前提。本文利用水体对激光信号的吸收特性, 设计了一种使用三维激光扫描检测水面目标的方法。该方法通过三维激光扫描获取水面漂浮物的点云, 然后通过条件滤波限定有效的点云集合, 并采用基于Kd树的最近邻聚类法划分水面目标物体, 最后使用基于轴向包围盒(Axis Aligned Bounding Box, AABB)计算目标相对激光雷达的位置。本文设计了三组水面目标物体检测试验, 数据表明本文方法可发现水面上漂浮的物体并能完成位置测定。
Abstract
The detection of surface target objects is a prerequisite for unmanned ship to identify surface objects or avoid obstacles. Based on the absorption characteristics of laser signal by water body, a method of detecting surface target by 3 d laser scanning is designed in this paper. The method obtains the point cloud of the surface float by 3D laser scanning, then defines the effective point cloud set by conditional filtering, and divides the water surface target object by the Kd-tree based nearest neighbor clustering method, finally uses the axial bounding box(Axis Aligned Bounding Box, AABB)calculates the position of the target relative to the lidar. Three sets of surface object detection experiments are designed in this paper. The data shows that the method can detect the floating objects on the water surface and can complete the position measurement.
向可均, 肖宇峰, 胡长青, 郑又能, 张华. 基于三维激光扫描的水面目标物体检测[J]. 应用激光, 2019, 39(4): 671. Xiang Kejun, Xiao Yufeng, Hu Changqing, Zheng Youneng, Zhang Hua. Water Surface Target Detection Based on 3D Laser Scanning[J]. APPLIED LASER, 2019, 39(4): 671.