Indoor feature point and structural line combination-based indoor SLAM (Simultaneous Localization and Mapping) method

Abstract

The present invention relates to an indoor feature point and structural line combination-based indoor SLAM (Simultaneous Localization and Mapping) method. The method includes the following steps that: S1, camera internal parameter calibration is performed; S2, feature points and structural lines are extracted from video frame image data acquired by a camera; S3, feature points and structural lines are tracked according to the obtained feature points and structural lines, and key frames are selected; S4, on the basis of the tracking information of the obtained feature points and structural lines, the space points and space lines of a surrounding environment are charted, and the positioning of a platform is optimized; S5, whether the motion trajectory of the platform forms a closed loop is judged, correct closed-loop key frames are obtained, and a global image pose and a map are optimized globally; and S6, a result is outputted. The method of the invention has the advantages of real-time performance and high efficiency. According to the method of the invention, matched feature points and structural lines are utilized to chart the pose of an image and the surrounding environment; loopback detection processing is performed; and therefore, when the structural lines are fully utilized to reduce drift errors, the good positioning result of the mobile robot platform and the structural features of the surrounding environment can be obtained with the loopback detection adopted.

Description

technical field [0001] The invention belongs to the field of photogrammetry and computer vision, and in particular relates to a visual SLAM algorithm combining indoor feature points and structural lines. Background technique [0002] With the development of computer vision and photogrammetry, graph optimization SLAM (Simultaneous Localization and Mapping) has attracted more and more attention from visual SLAM researchers. It introduces motion estimation and beam adjustment into SLAM. Motion estimation is the robot's The location and surrounding environment characteristics are solved as a global optimization problem. By extracting feature points and structural lines on the image, feature tracking is performed, and the observation error equation is established, and the optimal value is calculated by linear or nonlinear optimization to minimize the observation error value. robot position and surrounding environment features. Due to the time-consuming feature extraction, matchi...

AI Technical Summary

Problems solved by technology

[0005] Since feature extraction and matching consume a lot of calculations, image pixel-based visual SLAM has gradually attracted the attention of researchers. Image pixel-based visual SLAM algorithms directly use image grayscale information for image tracking without feature extraction. And description, directly use the gradient of the image for tracking and optimization, which can enhance the continuity of visual SLAM in areas with fewer features, but it completely relies on image pixels, and the effect is poor for areas with strong illumination changes, and only relies on images Pixel gradient, the calculated positioning and composition accuracy is relatively low

Images