摘要
船舶智能制造、数字孪生等技术的发展提出了从测量数据逆向重构船体结构模型的需求。由于三维激光扫描数据量巨大,点云特征识别和逆向建模是研究热点之一。由于常规的识别算法在非均匀分布和形状复杂的点云中存在明显的误差,因此提出基于表面法向一致性和高斯混合模型的板架结构识别方法,并应用于船舱三维点云的逆向建模。算法经过试验验证可行,能够可靠的实现船舱点云中平面特征提取,并应用于船舱逆向建模软件的实现。
Demand of hull structure reverse modeling is proposed by the development of intelligent ship building,digital twin and other emerging technologies.Due to the huge amount of 3D laser scanning data,point cloud feature recognition and reverse modeling have become research hotspots Since the conventional recognition algorithm has obvious errors in the point cloud with non-uniform distribution and complex point cloud,a plate frame structure identification method based on the normal consensus and Gaussian mixture model is proposed,and it is applied to the inverse modeling of the three-dimensional point cloud of the ship cabin.The algorithm is feasible after experimental verification,and can reliably extract plane features in the cabin point cloud,and is applied to the realization of cabin inverse modeling software.
作者
倪崇本
李志月
杨荣淇
NI Chongben;LI Zhiyue;YANG Rongqi(Shanghai Jiao Tong University,State Key Laboratory of Ocean Engineering,Shanghai 200240,China;Shanghai Jiao Tong University,School of Naval Architecture,Ocean and Civil Engineering,Shanghai 200240,China;Shanghai Merchant Ship Design and Research Institute,Shanghai 201203,China)
出处
《船舶工程》
CSCD
北大核心
2022年第2期123-127,共5页
Ship Engineering
基金
国防技术基础项目(JSJL2019206B001)
关键词
点云
逆向建模
法向一致性
高斯混合模型
极大似然估计
point cloud
reverse modeling
normal consensus
Gaussian mixture model
expectation maximum
