激光雷达水下障碍物剖面图像处理方法  

LiDAR profile image processing method for underwater obstacle

作　　者：阮英杰 贺岩[1,2,3] 吕德亮 侯春鹤 徐广袖 张超然 黄宜帆 郝歆珂 RUAN Yingjie;HE Yan;LV Deliang;HOU Chunhe;XU Guangxiu;ZHANG Chaoran;HUANG Yifan;HAO Xinke(Wangzhijiang Innovation Center for Laser,Aerospace Laser Technology and System Department,Shanghai Institute of Optics and Fine Mechanics,Chinese Academy of Sciences,Shanghai 201800,China;Key Laboratory of Space Laser Communication and Detection Technology,Shanghai Institute of Optics and Fine Mechanics,Chinese Academy of Sciences,Shanghai 201800,China;Center of Materials Science and Optoelectronics Engineering,University of Chinese Academy of Science,Beijing 100049,China;Naval Research Institute,Tianjin 300061,China)

机构地区：[1]中国科学院上海光学精密机械研究所空天激光技术与系统部王之江创新中心,上海201800 [2]中国科学院上海光学精密机械研究所空间激光传输与探测技术重点实验室,上海201800 [3]中国科学院大学材料与光电研究中心,北京100049 [4]海军研究院,天津300061

出　　处：《红外与激光工程》2024年第7期118-129,共12页Infrared and Laser Engineering

基　　金：国家重点研发计划项目(2022YFB3901705);国家实验室科技创新项目(LSKJ202201403)。

摘　　要：机载激光雷达已被广泛运用在水底地形测绘等领域,针对水下障碍物探测及识别的需求,基于线形扫描激光雷达的数据特点,提出将波形数据拼接为二维剖面图直观地反映回波能量的分布。使用上升沿线性近似的方法从回波提取水面斜距,并构建了基于机载激光雷达系统扫描结构的光线出射角度模型,通过对水面变形的校正实现了对水下障碍物回波轮廓的还原。使用Canny边缘检测算子提取相邻扫描线所成图像中高回波能量的边缘,最后通过Hu矩对比边缘结果间的相似度,构建障碍物识别判据并进行验证。该算法为机载激光雷达系统水下障碍物探测和识别的数据处理方式提供参考。Objective Detection of underwater obstacles represents a significant area of interest within marine detection technologies. Airborne LiDAR, an active detection modality, has found extensive application in domains including terrain surveying and underwater obstacle detection. The streak tube camera, by splitting laser light,considerably reduces the energy of emitted laser pulses for each channel. In contrast, single beam LiDAR detection capitalizes on the full energy of the emission pulse, facilitating deeper detection capabilities. LiDAR is particularly effective for detecting underwater obstacles in turbid water conditions. Considering the data characteristics of line scanning airborne LiDAR and the detection requirements, this study introduces an image processing approach inspired by the streak tube imaging system. The method involves splicing echo waveform data from each point on a scanning line to create a two-dimensional profile that directly illustrates the spatial distribution of echo energy, from which obstacle information is subsequently extracted and analyzed. An automated obstacle identification criterion is developed and validated. This research contributes to the refinement of data processing methods for underwater obstacle detection and identification using airborne LiDAR systems.Methods A sequence of LiDAR echo data strips is chronologically assembled;Each column corresponds to an echo waveform, with gray values representing the echo energy at each point. Rows, ordered from top to bottom,depict the amplitude at respective sampling moments. To address horizontal plane deformation due to scanning angle variations, the water surface slope distance is initially extracted from the waveform using the Linear Leading Edge Approximation(LLE) method. A model representing the emission angle of the laser light is then developed, based on the scanning architecture of the ocean LiDAR system. These components are integrated to pinpoint water surface points, facilitating the calculation of the discrepancy between

关 键 词：激光雷达 水下探测 图像处理 边缘检测 

分 类 号：TN249[电子电信—物理电子学]