Fast and robust visual odometry with a low-cost IMU in dynamic environments

Abstract

Purpose To realize stable and precise localization in the dynamic environments, the authors propose a fast and robust visual odometry (VO) approach with a low-cost Inertial Measurement Unit (IMU) in this study. Design/methodology/approach The proposed VO incorporates the direct method with the indirect method to track the features and to optimize the camera pose. It initializes the positions of tracked pixels with the IMU information. Besides, the tracked pixels are refined by minimizing the photometric errors. Due to the small convergence radius of the indirect method, the dynamic pixels are rejected. Subsequently, the camera pose is optimized by minimizing the reprojection errors. The frames with little dynamic information are selected to create keyframes. Finally, the local bundle adjustment is performed to refine the poses of the keyframes and the positions of 3-D points. Findings The proposed VO approach is evaluated experimentally in dynamic environments with various motion types, suggesting that the proposed approach achieves more accurate and stable location than the conventional approach. Moreover, the proposed VO approach works well in the environments with the motion blur. Originality/value The proposed approach fuses the indirect method and the direct method with the IMU information, which improves the localization in dynamic environments significantly.

TSG-SLAM: SLAM Employing Tight Coupling of Instance Segmentation and Geometric Constraints in Complex Dynamic Environments

“…Using information obtained from an IMU, a wheel odometer, and other sensors as prior motion knowledge for a camera can assist a system in segmenting dynamic targets. The SLAM system designed by Yao et al [ 17 ] includes tracking threads, feature extraction threads, and local mapping threads. One of the tasks of the tracking thread is to utilize the transformation matrix obtained from an IMU and combine it with the reprojection error to determine the dynamic nature of feature points.…”

An Adaptive ORB-SLAM3 System for Outdoor Dynamic Environments

“…A summary of related work with advantages and disadvantages is presented in Table 1 . In the study on semantic information-based for dynamic environments SLAM, Erliang Yao et al [ 9 ] used the transformation matrix of IMU and reprojection error to distinguish dynamic feature points.VDO-SLAM, proposed by Zhang et al [ 10 ] in 2020, uses semantic information to estimate accurate motion, but the system requires pre-processing datasets including instance-level semantic segmentation and optical flow estimation, so the system is computationally intensive. Zhang et al [ 11 ] used RGB-D cameras and a K-means clustering algorithm to distinguish foreground and background features in the bounding box.…”