Feature Point Tracking Method for Visual SLAM Based on Multi-Condition Constraints in Light Changing Environment

In scenes where there are lighting changes, localization may fail for visual SLAM due to feature point tracking failure. Thus, a method based on multi-condition constraints is proposed SLAM. The tracks the points of optical flow from aspects such as overall motion position points, descriptor grayscale information, and spatial geometric constraints. First, solve problem mismatch in complex environments, we propose removal that combines flow, descriptor, RANSAC. We eliminate incorrect matches layer by through these uniformity distribution image can then affect accuracy camera pose estimation, different also difficulty extraction. order balance quality extracted an adaptive mask homogenization adaptively adjusts radius according points. Experiments conducted EuRoC dataset show which integrates improved into tracking, exhibits robustness under various interferences blurring, unclear textures. Compared RANSAC method, reduce location error about 85% using dataset.