近景光学影像和三维激光扫描点云自动配准  

Automatic registration of close-range optical image and 3D laser scanning point cloud

作　　者：俞路路 李彩林[1] 王志勇[1] YU Lulu;LI Cailin;WANG Zhiyong(Shandong University of Technology,Architectural Engineering Institute,Zibo Shandong 255000,China)

机构地区：[1]山东理工大学建筑工程学院,山东淄博255000

出　　处：《激光杂志》2021年第6期92-96,共5页Laser Journal

基　　金：国家自然科学基金(No.41601496);国家自然科学基金(No.41701525)。

摘　　要：针对近景光学影像和三维激光扫描点云配准,采取先粗配准再精配准的策略,实现了近景光学影像和三维激光点云的自动配准。首先将具有一定重叠度的影像通过自由网平差获取相机内外方位元素,然后利用加入尺度因子的4PCS(4-points congruent sets)算法进行粗配准;最后基于空三点云与对应三维激光扫描点云切平面距离最小为原则,结合共线方程构建平差模型,实现精配准。粗配准为刚性配准精配准为非刚性配准,采用粗配准和精配准交替的策略,提高了精度配准精度。采用3组数据进行配准实验,与手工选取多对同名点进行了对比,结果表明粗配准算法精度较高,精配准后能取得理想效果。Aiming at the registration of close-range optical image and 3D laser scanning point cloud,the strategy of coarse registration and fine registration is adopted to realize the automatic registration of close-range optical image and 3D laser point cloud.The images with a certain degree of overlap are obtained by the free network adjustment to obtain the internal and external orientation elements of the camera,and then the 4PCS(4-points congruent sets)algorithm with scale factor is used for coarse registration;Based on the principle of minimum distance between the spatial three-point cloud and the corresponding three-dimensional laser scanning point cloud,the adjustment model is constructed combined with collinear equation to realize the precise registration.The coarse registration is rigid registration,and the fine registration is non rigid registration.The strategy of alternating coarse registration and fine registration improves the accuracy of registration.The results show that the coarse registration algorithm has higher accuracy and can achieve ideal results after precise registration.

关 键 词：粗配准 三维重建 精配准 

分 类 号：P237[天文地球—摄影测量与遥感] TN249[天文地球—测绘科学与技术]