Underground parking lot vehicle rapid repositioning method

Abstract

The invention provides an underground parking lot vehicle rapid repositioning method, and belongs to the field of automatic driving positioning and autonomous parking positioning. The rapid repositioning method comprises the following steps: aiming at a closed scene, collecting vehicle surrounding image information by a vehicle-mounted all-around view system, extracting external environment perception information, comparing the environment perception information with the data in the map, judging whether the environment perception information has a unique matching point matched with the data in the map, and if so, outputting the position of the matching point as a vehicle positioning result; if not, outputting a small number of matched candidate areas, comparing the environment perception information with data of the small number of candidate areas in the map, and performing iterative optimization output to obtain a vehicle positioning result. According to the vehicle rapid repositioning method, the cost of the field end equipment can be effectively saved, the accurate position of the vehicle can be converged more rapidly, even the repositioning result can be directly obtained, and the problem that rapid convergence cannot be realized due to the lack of semantic information in the existing technical scheme can be effectively solved.

Description

technical field [0001] The invention belongs to the technical field of automatic driving positioning and autonomous parking positioning, and in particular relates to a method for rapid repositioning of vehicles in an underground parking lot. Background technique [0002] In existing closed scenes such as underground parking lots, the following relocation methods are commonly used: [0003] 1. Communicate with the vehicle through the construction of UWB or laser equipment at the field end to inform the vehicle position. [0004] 2. By building a high-precision map or dense laser point cloud, by matching the structured road markers in the scene or matching point cloud features, multiple candidate positions are given and converged to a suitable position in the motion through the filter. [0005] However, it needs a lot of construction and follow-up maintenance costs to communicate with the vehicle through the construction of UWB or laser equipment at the field end to inform th...

AI Technical Summary

Problems solved by technology

[0005] However, it needs a lot of construction and follow-up maintenance costs to communicate with the vehicle through the construction of UWB or laser equipment at the field end to inform the vehicle position

By building a high-precision map or a dense laser point cloud, by matching the structured road markers in the scene or matching the point cloud features, multiple candidate positions are given and converged through the filter to a suitable position in the movement. At the same time, the semantic information of road markers has a high repetition rate in underground parking lots, and the convergence speed is very slow in large-scale scenes, which affects user experience, and point cloud features require a long period of time for matching and storage space There is a greater demand

Images