基于RGB信息的激光点云数据配准与分割精简算法

An Algorithm of Laser Point Cloud Data Registration and Segmentation Based on RGB Information

针对现有激光点云配准和分割算法存在的配准效率低、精度差及对激光点云数据集合精简效果差等不足,提出一种基于RBG信息的配准与分割精简算法。先借助RGB相机重新标定三维激光图像,为三维激光点云提供准确的参量,通过RBG坐标转换提升对点云集合的定位精度。基于四元素法配准激光点云的图像信息,改善配准精度和收敛速度;在空间内分割独立的包围体,形成多个小格栅,并提取RGB色彩特征,实现对冗余点云数据的精简。实验结果显示,RGB信息算法的配置精度和配置效率均优于传统算法,能够获得较好的数据精简效果。

Aiming at the shortcomings of existing laser point cloud registration and segmentation algorithms,such as low registration efficiency,poor accuracy and poor reduction effect on laser point cloud data sets,a registration and segmentation reduction algorithm based on RBG information is proposed.First,the RGB camera is used to recalibrate the 3D laser image to provide accurate parameters for the 3D laser point cloud,and the positioning accuracy of the point cloud set is improved through RBG coordinate conversion.The image information of laser point cloud is registered based on the four element method to improve the registration accuracy and convergence speed.Finally,separate bounding volumes are segmented in space to form multiple small grids and extract RGB color features to simplify redundant point cloud data.The experimental results show that the RGB information algorithm is superior to the traditional algorithm in configuration accuracy and configuration efficiency,and can achieve better data reduction effect.