区域增长与RANSAC模型结合的机载点云平面分割方法

SEGMENTATION METHOD OF AIRBORNE POINT CLOUD PLANE BASED ON REGIONAL GROWTH AND RANSAC MODEL

针对现有的建筑物点云分割数据效率低、抗噪性弱等问题,提出了一种区域增长与RANSAC模型结合的点云平面分割方法,该方法以种子点的曲率、法向量及其邻域点到法平面的欧氏距离为生长约束条件,通过不断扩展种子面提取符合平面模型的初始样本点集,然后采用基于RANSAC平面拟合的稳健参数估计方法估计出高精度的平面参数。试验证明该方法具有可行性。

Aiming at the problems of low efficiency and low noise resistance of existing building point cloud segmentation data,a point cloud plane segmentation method based on region growing and RANSAC model fitting was proposed.The method took the curvature of seed points,normal vector and euclidean distance from its neighborhood points to normal plane as growth constraints,and extracted the initial data in line with the plane model by continuously expanding the seed surface.Then,the robust parameter estimation method based on RANSAC plane fitting was used to estimate the plane parameters with high accuracy.The experimental results show that the method is feasible.