摘要
针对水火弯板机检测系统中船舶外板三维点云数据自动提取过程存在边缘噪点、板下贴合垫木识别效率低以及外板边缘拟合等问题,提出一种优化DBSCAN聚类算法,首先根据现场加工环境精简点云,利用网格划分来建立外板点云拓扑模型,然后根据DBSCAN密度聚类算法搜索出外板点云,最后采用最小二乘法进行边缘拟合。结果表明,该算法能有效识别外板边缘,提取出外板点云。
Detection system for plate bending by line heating in the ship's outer edge three-dimensional automatic extraction of point cloud data, sheet laminating wood recognition under noisy fitting problems such as low efficiency as well as the outer edge. A optimization clustering algorithm DBSCAN is presented, based on point cloud processing environment to streamline, using mesh points to establish outer topological model, then search out based on density clustering algorithm DBSCAN Board point cloud, using least square fitting the edge. Experimental results show that the algorithm can identify the outer edge extraction points out plate.
出处
《船舶工程》
北大核心
2015年第8期74-78,共5页
Ship Engineering
基金
高性能大曲率复杂曲面船体外板成型智能机器人(2012A080103007)
关键词
船体外板
三维点云数据
DBSCAN聚类
水火弯板
点云提取
模式识别
hull of plate
3D of point cloud data DBSCAN clustering
plate bending by line heating
point cloud extraction
pattern recognition