Urban structured scene pure vision positioning method and device for automatic driving vehicle

Abstract

The invention discloses an automatic driving vehicle city structured scene pure vision positioning method and device, and the method comprises the steps: setting two parallel threads: a vision SLAM positioning thread which is used for obtaining the current pose of a vehicle in real time through employing a local map-based ORB feature point method; the offline semantic map matching and positioning thread comprises the following steps: 1) training and detecting a structured semantic marker in an original image acquired by a camera; 2) acquiring two-dimensional information of the marker in an image coordinate system according to ground prior; (3) time domain tracking is conducted on the markers, when the number of the markers detected and obtained in the same frame is not smaller than a preset number, a two-dimensional topological graph of the vehicle and the markers is constructed, and the topological edge weight of the two-dimensional topological graph is determined by the reliability of the markers associated with the current frame; 4) performing map search matching according to the current position of the vehicle and the two-dimensional topological graph, and recovering the current two-dimensional pose of the vehicle; and 5) uploading the information of the moving trajectory and the marker information to a cloud end, and updating an offline map semantic information base.

Description

technical field [0001] The present invention relates to the technical field of automatic driving, in particular to a method and device for purely visual positioning of urban structured scenes of automatic driving vehicles based on prior maps. Background technique [0002] AV (Autonomous vehicles; self-driving vehicles) can calculate their own pose through on-board sensors, and combine pose information for autonomous planning and horizontal and vertical control, and positioning methods must be used. Positioning methods are roughly divided into absolute positioning based on GNSS (Global Navigation Satellite System) and relative positioning based on environment-aware sensors such as cameras, lidars, and IMUs (Inertial Measurement Units). [0003] Absolute positioning usually depends on the availability of GNSS, GNSS includes such as GPS (Global PositionSystem; Global Positioning System), Glonass (global navigation satellite system; Glonass system), Beidou satellite navigation s...

AI Technical Summary

Problems solved by technology

Most of the existing technologies solve this problem by associating with the global map. However, the positioning problem under long-term and complex urban working conditions without GPS signals still needs to be solved urgently.

Images