Point cloud data processing method and device

Abstract

The invention discloses a point cloud data processing method and device. The method comprises the following steps: acquiring a point cloud distribution image; grouping the point cloud data; determining a reference point set from the point cloud data set; determining a ground reference line according to the reference point set; and judging whether the point cloud corresponding to the ground reference line in each group belongs to the ground or not according to the ground reference line. According to the method, the purpose of identifying whether the point cloud belongs to the ground or not is achieved, so that the robot can efficiently distinguish obstacles and the ground in the sight line range, such as flat ground, small gullies and slopes which do not affect trafficability, and a front-end foundation is laid for the robot to flexibly and autonomously avoid obstacles in extremely complex scenes. According to the technical scheme of the invention, effective identification of the ground point cloud can be realized by using an image shot by a common depth camera, and an expensive industrial-grade camera is not needed, so that the scheme cost is reduced, and the productization and mass production of the scheme are facilitated.

Description

technical field [0001] The present application relates to the technical field of image processing, in particular, to a point cloud data processing method and device. Background technique [0002] In the field of point cloud image processing, the existing ground removal methods generally include direct cutting method and normal vector comparison method. The direct resection method has relatively high requirements on hardware and environment, and requires a very accurate camera with low noise. It also has very high requirements on the flatness of the ground. If there are flat land and slopes in the map, this method is completely unsuitable. The normal vector comparison method compares the point cloud normal vector with the normal vector threshold to obtain a point cloud that meets the ground requirements, and then deletes the point cloud data to obtain the point cloud data after the ground is removed. This method also has relatively high requirements on the camera, and too mu...

AI Technical Summary

Problems solved by technology

This method also has relatively high requirements on the camera, and there should be no too much noise, because the noise is too large and the normal vector is prone to deviation

If the ground captured by ordinary consumer-grade cameras is incomplete, the error in calculating the normal vector will be relatively large, and this method consumes a lot of computing resources during data processing, and cannot be commercialized

The above two methods have high requirements on the camera and hardware, but the high-precision camera is very expensive

The price cannot meet the requirements of commercial mass production

Images