RGB-D visual SLAM method based on dot-line features

Abstract

The invention discloses an RGB-D visual SLAM method based on dot-line features. The RGB-D visual SLAM method comprises the following steps: S1, calibrating internal parameters of an RGBD camera; s2, performing line feature extraction; s3, acquiring an initial transformation pose of the camera; s4, performing linear projection from the key frame to the current frame; s5, forming an error function of the visual odometer module; s6, performing iterative optimization of the initial transformation pose of the current frame of camera and pose solution of the current frame of camera; s7, judging thekey frame; s8, performing local BA optimization; s9, performing loop detection; s10, performing repositioning; s11, constructing a map; s12, outputing a result. According to the SLAM method provided by the invention, the RGB-D camera is used as image acquisition equipment, points and edge lines of an object in a scene are used as positioning features to realize SLAM, and the method has relativelygood accuracy in challenging scenes such as low semantics and low texture.

Description

technical field [0001] The invention relates to an RGB-D visual SLAM method based on point-line features, and belongs to the technical field of visual SLAM (synchronous positioning and map construction). Background technique [0002] SLAM (simultaneous localization and mapping, synchronous positioning and map construction) can be summarized as: a certain mobile device (such as a robot, a drone, a mobile phone, etc.) starts from an unknown location in an unknown environment, and passes through sensors (such as Laser radar, camera, etc.) observe and position its own position, attitude, and motion trajectory, and then perform incremental map construction according to its own position, so as to achieve the purpose of simultaneous positioning and map construction. The development of SLAM technology has a history of more than 30 years, involving many technical fields, and has become a popular research direction in the field of robotics and computer vision. [0003] Generally spea...

AI Technical Summary

Problems solved by technology

[0005] In view of the above-mentioned problems existing in the prior art, the purpose of the present invention is to provide a RGB-D visual SLAM method based on point-line features, so as to solve the problem that the current visual SLAM technology has low precision in low semantics, large illumination changes, and dynamic environments. The problem

Images