Multi-sensor fusion road extraction and indexing method based on global and local grid maps

Abstract

The invention discloses a multi-sensor fusion road extraction and indexing method based on global and local grid maps. Data storage and preprocessing are carried out by adopting global and local gridmaps; the real-time position and posture of an unmanned vehicle are obtained through a differential positioning system. High-precision road boundaries lane lines and road sign positions are obtainedthrough laser radar point cloud related data stored in the grid map and projected into the monocular vision image, the lane line type is judged based on a deep learning method, road sign semantic information is recognized, and therefore a high-precision road map with semantic information is obtained. Efficient storage and indexing are realized in a mode of establishing multiple global sections androad sets.

Description

technical field [0001] The invention belongs to the technical field of automatic driving, and relates to an efficient and high-precision semantic road map construction method and a visualization method thereof. Background technique [0002] High-precision maps are of great significance in the field of autonomous driving. A high-precision map, also called a high-resolution map, is a map specially designed for driverless driving. Compared with traditional maps, high-precision maps can not only provide road-level navigation information, but also lane-level positioning and navigation information. [0003] In the current high-precision map drawing scheme, a large number of steps such as laser point cloud conversion, splicing, labeling and indexing are required, and the process is very complicated and requires professional technicians. This makes high-precision maps expensive, highly specialized, and inefficient in production. In addition, in the usual high-precision map produc...

AI Technical Summary

Problems solved by technology

[0003] In the current high-precision map drawing scheme, a large number of steps such as laser point cloud conversion, splicing, labeling and indexing are required, and the process is very complicated and requires professional technicians

This makes high-precision maps expensive, highly specialized, and inefficient in production

In addition, in the usual high-precision map production process, there is no original semantic information, and only manual labeling of some lane lines and road signs, etc., is inefficient and the semantic information is single

At the same time, the storage and indexing methods of the existing high-precision maps are complex, which also rely on a large amount of labor, and the analysis methods are complicated, which is not conducive to the visual display during vehicle driving

Images