Panoramic vision SLAM method based on multi-camera cooperation

Abstract

The invention discloses an SLAM method based on multi-camera cooperation. 360-degree observation of a detection environment is realized through a plurality of cameras which supplement visual fields mutually; each camera independently acquires and analyzes data; meanwhile, in the processes of map construction and motion tracking, the map is tracked; the effective camera collaboratively completes the tasks of generation of shared map points with scale information and joint pose estimation; map construction and high-precision positioning tasks of the panoramic SLAM system can be completed throughcooperation of imaging and structural characteristics of multiple cameras. The problem that the perception visual field of the existing pure visual SLAM system is limited, the positioning precision of a weak texture environment is poor; the map scale information lacks, illumination resistance and shielding capability are weak and the like are solved, meanwhile, a plurality of cameras are independent of one another, the normal work of the system cannot be influenced when any camera goes wrong, the system has certain robustness on the conditions of object shielding, direct sunlight, lens damage and the like, and a technical guarantee is provided for autonomous navigation of the unmanned platform.

Description

Technical field [0001] The invention belongs to the technical field of computer vision positioning, and specifically relates to a panoramic vision SLAM method based on multi-camera collaboration. Background technique [0002] With the rapid development of artificial intelligence technology, research in fields such as unmanned driving, virtual reality, and face recognition has become a current hot spot. Among them, unmanned driving has achieved lane-level positioning effects in some urban environments with known maps, but driving on unstructured roads with unknown environments, and cannot use signal source positioning sensors such as GPS, Beidou, Galileo, etc. Under the circumstances, how to achieve autonomous map construction and precise positioning is one of the difficulties in this field. [0003] Simultaneous Localization and Mapping (SLAM) refers to the unmanned platform using unmanned sensors such as cameras, Lidar, odometer, inertial sensors and GPS and Beidou with signal so...

AI Technical Summary

Problems solved by technology

[0006] In view of this, the purpose of the present invention is to provide a panoramic visual SLAM method based on multi-camera collaboration, which can solve the problems of the existing pure visual SLAM system, such as limited perceptual field of view, poor positioning accuracy in weak texture environments, lack of map scale information, and anti- Problems such as poor lighting and occlusion capabilities provide strong technical support for autonomous navigation of unmanned platforms in unstructured environments

Images