基于三维成像声呐的无人艇水下目标检测  

Underwater Target Detection for Unmanned Surface Vessel Based on Three-Dimensional Imaging Sonar

作　　者：杨一 黄斌[1] 周新[1] 崔化超 YANG Yi;HUANG Bin;ZHOU Xin;CUI Huachao(The 28th Research Institute of China Electronics Technology Group Corporation,Nanjing 210023,China)

机构地区：[1]中国电子科技集团公司第二十八研究所,南京210023

出　　处：《指挥信息系统与技术》2024年第2期76-82,共7页Command Information System and Technology

摘　　要：针对无人艇水下目标探测过程中声呐图像受干扰多、单个目标可能分裂为多个亮斑等问题,开展了基于三维成像声呐的无人艇水下目标检测技术研究。首先,考虑水面无人艇水下目标检测跟踪识别的需要,建立了三维点云图像模型;然后,提出了基于随机抽样一致性(RANSAC)平面拟合的海底检测算法,以完成三维点云图中海底平面提取与分离,通过基于欧式聚类的三维点云检测技术和多特征数据关联技术等手段提高了目标的检测概率;最后,在湖上进行了试验,试验结果验证了该算法对水下目标检测识别的有效性。Aiming at problems about frequent interference in sonar images and the possibility of a single target splitting into multiple bright spots during the underwater target detection process of unmanned surface vessel(USV),research on underwater target detection technology for USV based on three-dimensional(3D) imaging sonar is carried out.Firstly,considering the need of detection,tracking and recognition to underwater targets of USV,a 3D point cloud image model is established.Then,a seabed detection algorithm based on random sample consensus(RANSAC) plane fitting is proposed to complete the extraction and separation of seabed planes in the 3D point cloud image.By using technologies of 3D point cloud detection based on Euclidean clustering and multi-feature data association,the probability of target detection is enhanced.Finally,experiment in a lake is carried out.Experimental results verify the effectiveness of the algorithm for underwater target detection and recognition.

关 键 词：三维点云图 随机抽样一致性(RANSAC)平面拟合 海底检测 欧式聚类 多特征数据关联 

分 类 号：TB566[交通运输工程—水声工程]