基于RGB-D SLAM的三维点云地图融合算法

3D Point Cloud Map Merging Algorithm based on RGB-D SLAM

针对地图融合算法初始化状态设定复杂的问题,提出一种判断地图重叠区域,同时提供最近点迭代(Intrative Closest Point,ICP)算法初始变换矩阵的算法。在RGB-D SLAM建图的基础上,通过特征点匹配和对极约束寻找地图重叠区域,并计算地图之间的初始变换矩阵,经过ICP算法迭代后得到地图之间的精确变换矩阵。为了验证算法精度和实际效果,在公开数据集和实际环境中分别运行该算法,结果表明该算法拥有较高的精度。

To solve the problem of complex initialization state setting of map merging algorithm,an algorithm is proposed to determine the overlapping area of maps and provide the initial transformation matrix of ICP algorithm.On the basis of RGB-D SLAM mapping,the overlapped area of maps is found by feature point matching and epipolar constraint,and the initial transformation matrix between maps is calculated.After iteration of ICP algorithm,the accurate transformation matrix between maps is obtained.In order to verify the accuracy and actual effect of the algorithm,the algorithm is run in public datasets and real environment respectively.The results show that the algorithm has high accuracy.