DT-SLAM: Dynamic Thresholding Based Corner Point Extraction in SLAM System

Abstract

Visual localization estimation is highly depended on the quality of video frames or captured images. Estimation quality may be affected by the poor visibility, low background texture and overexposure. Low quality frames with blurred edges and poor contrast pose tremendous difficulties for corner points detection in SLAM impacting the overall accuracy of estimation. This paper introduces DT-SLAM, a dynamic self-adaptive threshold (DSAT) approach for ORB corner points extraction in FAST to improve SLAM's localization performance. The proposed method replaces the existing static threshold-based ORB extraction approach, enabling improved performance in complex real-world scenes. In addition, this study introduces a threshold switching mechanism (TSM) to replace the existing SLAM's frame-level and celllevel thresholds for corner points extraction. The proposed DT-SLAM approach is validated using the TUM RBG-D and EuRoC benchmark datasets for location tracking performances. The results indicate that the proposed DT-SLAM (mean RM SE = 0.0807 (EuRoC); mean RM SE = 0.0153 (TUM)) outperforming the current state-of-the-art ORB-SLAM3 (mean RM SE = 0.1899 (EuRoC); mean RM SE = 0.0252 (TUM)).

An Improved ORB Feature Extraction Algorithm Based on Enhanced Image and Truncated Adaptive Threshold

“…In recent years, scholars have mainly made two improvements to address the illumination variation problem in the ORB algorithm: the first is to improve the calculation method of adaptive thresholds [17]- [19], and the second is to expand the application scope of adaptive thresholds [20], [21].…”

Dynamic Object Detection and Tracking in Vision SLAM

“…Vision-based synchronous localization and map construction technology is a key technology in the field of intelligent robot research [7][8][9]. Among them, vision SLAM technology with the camera as the main sensor has gradually become a hot spot of research in industry and academia with the advantages of low cost, easy installation, and a wide range of applications [10][11]. However, the traditional vision SLAM technology is based on the assumption of a static environment, and dynamic object detection in the real scene is always the key and difficult problem to solve by vision SLAM technology [12][13][14].…”