基于多结构快速生成算法的建筑物平面提取-《集美大学学报（自然科学版）》

文章信息/Info

Title:: The Extraction of the Plane of Building Based on Accelerating Hypothesis Generation for Multi-structure

Author(s):: XIE Jiao; WU Yun-dong; CHEN Shui-li; School of Science，Jimei University，Xiamen 361021，China

Keywords:: Point cloud data; plane extraction; hypothesis generation; RANSAC

摘要:: 针对RANSAC算法在多结构数据集中提取平面点时存在的不足，提出了基于多结构快速生成算法的点云平面提取的新算法.该算法在随机产生一组平面模型之后，通过每个点相对于模型的残差排序信息，计算条件内点概率分布，然后利用得到的内点先验分布概率指导模型采样.实验结果表明，该算法能准确地检测出点云数据中的平面，相比RANSAC算法具有更好的采样效率.

Abstract:: This paper analyzed the deficiency of RANSAC that applying to the mutil-structure data set,then a plane fitting algorithm based on accelerating hypothesis generation for mutilstructure algorithm is proposed.Firstly,the proposed method uses random sampling to generate a set of models.For each model,all pixels have been sorted by calculating their residuals.As a consequence,the sampling process is guided by the inlier conditional probability distribution,which is calculated by residual sorts of each model.The experimental results show that the proposed method can accurately extract planes from LIDAR point cloud and it can obtain higher efficiency of sampling than the RANSAC.