Three-dimensional high-precision map generation method and device

Abstract

The invention discloses a three-dimensional high-precision map generation method and device which are used for solving the problems that an existing three-dimensional high-precision map generation method is high in cost and low in efficiency. The three-dimensional high-precision map generation method comprises: collecting environment information around a vehicle body in the moving process of a vehicle in real time, and the environment information comprising environment point cloud data; for each frame of environment point cloud data, executing the following operations: carrying out object detection and semantic segmentation on the environment point cloud data according to a preset neural network model, determining movable obstacle point cloud data, and removing the movable obstacle point cloud data from the environment point cloud data; carrying out point cloud registration on the environment point cloud data without the movable obstacle point cloud data and the stored environment point cloud data with the movable obstacle point cloud data removed in the previous frame according to a preset algorithm; and completing one-time incremental updating of the three-dimensional high-precision map according to the environment point cloud data subjected to the point cloud registration.

Description

technical field [0001] The invention relates to the technical field of map construction, in particular to a method and device for generating a three-dimensional high-precision map. Background technique [0002] In the related fields of unmanned driving technology, the three-dimensional high-precision map plays an important role. It contains a large amount of semantic information and driving assistance information. Unmanned vehicles quickly and accurately identify obstacles such as pedestrians and vehicles, and provide accurate current positions for unmanned vehicles. [0003] The research of 3D high-precision maps involves the collection, processing and storage of a large amount of point cloud data, and involves the automatic extraction of semantic information. In the prior art, common 3D high-precision maps are generated offline, which requires expensive data collection vehicles to identify the elements in the collected point cloud data through machines and manual semi-aut...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problems of high cost and low efficiency of the existing 3D high-precision map generation method, the embodiment of the present invention provides a 3D high-precision map generation method and device

Images