摘要
针对ICP算法在初始位姿差、部分数据丢失和噪声干扰情况下鲁棒性差、配准精度低的问题,提出一种基于统计局部特征描述与匹配的点云配准算法。首先,用点云局部密度、点云拟合平面距离方差、高斯曲率和平均曲率构建一个四维的统计局部特征描述符,准确地描述查询点的局部特征;然后,通过点对间的特征差异进行对应点匹配,剔除错误点对,解决点云部分数据缺失和噪声干扰的问题;最后,使用平均匹配距离作为度量改进ICP算法,对点云进行配准,解决初始位姿较差时配准精度低的问题。实验结果表明,该算法在初始位姿差、部分数据丢失和噪声干扰情况下的配准精度提高至少1个量级,配准速率也有较大提升,在鲁棒性和配准精度方面均表现出明显优势。
A point cloud alignment algorithm based on statistical local feature description and matching is proposed to address the problems of poor robustness and low alignment accuracy of the ICP algorithm in the presence of poor initial positional,partial data loss,and noise interference.Firstly,a 4-dimensional statistical local feature descriptor is constructed using point cloud local density,point cloud fitting plane distance variance,Gaussian curvature,and mean curvature to accurately describe the local features of the query points.Then,the corresponding points are matched by the feature difference between point pairs to eliminate the wrong point pairs and solve the problems of missing data and noise interference in part of the point cloud.Finally,the mean matching distance(MMD)is used as a metric to improve the alignment accuracy.MMD is a metric to improve the ICP algorithm to align the point clouds and solve the problem of low alignment accuracy when the initial poses are poor.The experimental results show that the algorithm improves the alignment accuracy by at least one order of magnitude and saves the alignment time in the case of poor initial poses,partial data loss,and noise interference,showing significant advantages in terms of robustness and alignment accuracy.
作者
王鑫淼
李新春
陶志勇
WANG Xinmiao;LI Xinchun;TAO Zhiyong(School of Electronic and Information Engineering,Liaoning Technical University,Huludao 125100,China)
出处
《液晶与显示》
CSCD
北大核心
2024年第1期89-99,共11页
Chinese Journal of Liquid Crystals and Displays
基金
2022年辽宁省应用基础研究计划(No.2022JH2/101300274)。
关键词
点云配准
特征描述
特征匹配
平均匹配距离
迭代最近点
point cloud registration
feature description
feature matching
mean match distance
ICP
