Point cloud processing method, device and equipment and storage medium

Abstract

The invention relates to the technical field of point cloud maps, and provides a point cloud processing method, device and equipment and a storage medium. The method comprises the steps: obtaining a plurality of point cloud tracks which are divided from an original point cloud track and distributed at different heights; based on an initial section point cloud trajectory determined in the multiplesections of point cloud trajectories, obtaining candidate section point cloud trajectories in the multiple sections of point cloud trajectories, wherein the candidate section point cloud trajectoriesand the projections of the initial section point cloud trajectories in the height direction do not have point cloud projection overlapping; determining a splicing section point cloud track communicated with the starting section point cloud track in the candidate section point cloud tracks; and forming a target point cloud trajectory based on the spliced section point cloud trajectory and the initial section point cloud trajectory, thereby ensuring continuity of subsequent semantic annotation processing on the target point cloud trajectory, and ensuring accuracy of semantic annotation on the point cloud trajectory on an annotation plane.

Description

technical field [0001] The present application relates to the technical field of point cloud maps, in particular to a point cloud processing method, device, computer equipment and storage medium. Background technique [0002] Point cloud map processing involves semantic labeling of point cloud trajectories. Among them, the point cloud trajectory formed by the point cloud may span different heights in three-dimensional space, and there are some problems in the scheme of marking such point cloud trajectory provided by the current technology. Specifically, in one of the labeling methods, the point cloud trajectories distributed at different heights are usually divided into multiple point cloud trajectories according to their distribution heights. For two point cloud trajectories with large differences in height distribution, the In other words, even if they are connected to each other, they will be semantically annotated at their respective heights because they are distributed...

AI Technical Summary

Problems solved by technology

However, if another method is used, that is, to semantically label the point cloud trajectory in 3D space directly according to the connectivity, then it will be impossible to map the connected points on the labeling plane due to the problem of overlapping projections of the point cloud trajectory on the labeling plane. Accurate labeling of cloud trajectories

Images