Vision-based automatic guided vehicle positioning method

Abstract

The invention discloses a vision-based automatic guided vehicle positioning method, including: collecting binocular images in front of the AGV and performing feature point matching, and calculating the depth of feature points for successful matching; using the left-eye image after discarding unmatched feature points as the current Picture frame, and match its feature points with the feature points of the previous key frame in the map, and use the ICP algorithm to solve the current pose of the left-eye camera according to the matching result; according to the current pose of the left-eye camera and the distance between the AGV and the left-eye camera Calculate the current position of the AGV based on the pose transformation relationship; whether there is a QR code positioning information input, yes, set the current picture frame as the current key frame and perform global optimization; no, judge whether the current picture frame meets the preset conditions, and satisfy Then set the current picture frame as the current key frame and do local sliding window optimization; if not satisfied, re-acquire the image; use the key frame and its feature points that have been optimized globally or locally sliding window to update the map and enter the next positioning.

Description

Technical field [0001] The present invention relates to the field of robots, and specific relates to a visual automatic guidance vehicle positioning method. Background technique [0002] Automated guidance car, also known as Automated Guided Vehicle, AGV), refers to an automatic guidance device such as magnetic strip, track or laser, travels along planned paths, battery powers, and equipped Safety protection and drone automated vehicles of various auxiliary institutions (such as moving, assembly agencies). At the monitoring and task schedule of the control center, the AGV can accurately walk according to the prescribed path, and after reaching the task specified position, complete a series of job tasks. [0003] The current AGV positioning navigation mode mainly has electromagnetic navigation, magnetic stripe navigation, inertial positioning navigation, two-dimensional code positioning navigation, laser positioning navigation and other ways. [0004] Electromagnetic navigation ne...

AI Technical Summary

Problems solved by technology

[0009] The main purpose of the present invention is to overcome the deficiencies of the existing positioning methods, and propose a vision-based automatic guided vehicle positioning method. During the positioning process, the cumulative error of the visual odometer positioning is continuously corrected by using the two-dimensional code positioning technology, thereby Guaranteed high-precision real-time positioning

Images