3D-LiDARを用いた階段上の人物および物体の点群検出手法の開発

抄録

In this research, we developed a method that detects people and objects on stairs using 3D point cloud information acquired from 3D-LiDAR. In this method, the size of the cluster is determined using a unique bounding box. The base bottom surface is defined by the OBB of the XY coordinate information from each cluster generated by DBSCAN, and the height is defined by the maximum and minimum coordinates of each cluster. In this method, RANSAC, wall recognition, noise removal, etc. are performed from the acquired point cloud, and the target point cloud is extracted using the two-dimensional histogram of the point cloud. The side of the staircase is removed by dividing the case by the variance of the inner product value of the normal vector, and the detection of the 3D point cloud of human and non-human objects is realized.