摘要
针对3D复杂点云模型,提出一种基于能量值的半刚性配准算法。该算法首先对点云进行基于法向量夹角距离的分割处理从而生成多个刚性子点云。通过整数规划法计算子点云间的配准能量从而推测配准的初始3D变换矩阵。基于初始矩阵,点云的最终配准结果由子点云间刚性配准方法 ICP实现。实验结果证明该方法可有效应用于模拟场景以及真实复杂场景的配准应用中。
We proposed an energy value-based semi-rigid registration method for 3D complex point cloud model. First the method segments the point cloud based on distance of normal vector angle so as to generate a couple of rigid sub-point clouds. Then it calculates the alignment energy between sub-point clouds with integer programming method so that to speculate on the initial 3D transformation matrix of registration.Based on initial matrix,the finial registration result is implemented by rigid alignments method ICP among sub-point clouds. Experimental results demonstrated that the method can be effectively applied to the registration applications in simulative scenes and real complex scenes.
出处
《计算机应用与软件》
CSCD
2016年第3期184-187,192,共5页
Computer Applications and Software
基金
国家自然科学科学基金项目(61300082)
辽宁省教育厅科学研究项目(L2014575)
大连东软信息学院博士启动基金项目(ZX2014KJ003)
关键词
半刚性配准
点云分割
点云能量
整数规划法
Semi-rigid registration
Point cloud segmentation
Point cloud energy
Integer programming