Track edge matching method for incremental map construction, electronic equipment and storage medium

Abstract

The invention provides a track edge matching method for incremental map construction, electronic equipment and a machine readable storage medium. According to the method, after initial point cloud registration is carried out on candidate edge matching point pairs to obtain pose changes of successfully matched laser point clouds to be put in storage, the pose changes of the successfully matched laser point clouds to be put in storage are transmitted to the laser point clouds to be put in storage which fail to be matched within a set range around the laser point clouds in a relative pose transmission processing mode; the method comprises the following steps: performing point cloud registration to adjust a pose initial value of a matching failure position, then performing secondary point cloud registration, and finally performing track edge matching optimization on a set by using successfully matched point clouds obtained by two times of point cloud registration. By transmitting the pose change of the successfully matched position to the failed matching position in the local range, the pose initial value adjustment can be effectively performed on the local track, so that the success rate of laser point cloud registration is effectively improved, the success rate of track edge matching is improved, and a high-precision reference track is provided for incremental construction of a high-precision map.

Description

technical field [0001] The invention relates to the technical field of automatic driving, in particular to a track edge method for incremental map construction, electronic equipment and a machine-readable storage medium. Background technique [0002] Due to the need to cover large-scale map data at the city level, high-definition maps (High Definition Maps, referred to as HD Maps) for autonomous driving often require multiple acquisitions and incremental construction. In the process of incrementally building a high-precision map, in order to ensure the global consistency of the map, it is necessary to process the map data of adjacent road sections. The success of edge joining depends on the overlap rate of the laser point cloud, the relative and absolute accuracy of the initial trajectory, etc. In complex urban scenes such as elevated roads, urban canyons, etc., due to the decline in the accuracy of integrated navigation systems that rely on GNSS (Global Navigation Satellit...

AI Technical Summary

Problems solved by technology

In complex urban scenes such as elevated roads, urban canyons, etc., due to the decline in the accuracy of the integrated navigation system that relies on GNSS (Global Navigation Satellite System, Global Navigation Satellite System) to obtain absolute positions, track bordering is more difficult than in high-speed scenes

Images