Vision-based topological map generation method

Abstract

The invention discloses a vision-based topological map generation method, which comprises the following steps: acquiring image information, and extracting image frames at a predetermined frequency based on the image information; obtaining a key frame based on image frame preprocessing, judging whether a loop is generated currently or not, if so, correcting a current existing topological map, and if the key frame is a first frame of image in the mapping operation process, setting a topological node at the position where the key frame is located; discrimination modes of other topological pointscomprise two modes: two modes: 1, comparing image information with an image of a newly added topological node in an existing topological map, judging whether the descriptor similarity of the two images is lower than a first threshold value or not, if the descriptor similarity is lower than the first threshold value, adding the position of an image of a key frame into the topological map as a new topological node, and storing a current frame; and 2, judging whether the angle information of a plurality of continuous frames is greater than a second threshold value relative to the previous topological node, if so, adding the position of the image of the key frame into the topological map as a new topological node, and storing the current frame.

Description

Technical field [0001] The present invention belongs to the field of computer vision technology, in particular to a vision-based topological map generation method. Background technique [0002] Unmanned systems, such as self-driving unmanned vehicles and unmanned aerial vehicles, are systems that use the built-in sensors of the system to acquire and describe the driving status of the system, and use the acquired information to complete autonomous navigation tasks. Sensors used in general unmanned systems include lasers, radars, cameras, IMUs, etc. Due to the high prices of lasers and radars, and the relatively low prices of cameras, navigation technology based on computer vision technology has been widely studied and applied. [0003] Simultaneous Localization and Mapping (SALM) technology refers to a system equipped with specific sensors to estimate its own state through motion and build a map that includes surrounding environment information during the motion. When we mainly us...

AI Technical Summary

Problems solved by technology

However, in the existing methods, the generation of topological maps lacks effective logic, and the efficiency of the generated nodes for space representation is relatively low.

Images