ULG-SLAM: A Novel Unsupervised Learning and Geometric Feature-Based Visual SLAM Algorithm for Robot Localizability Estimation

Yihan Huang,Fei Xie,Jing Zhao,Zhilin Gao,Jun Chen,Fei Zhao,Xixiang Liu

DOI: https://doi.org/10.3390/rs16111968

IF: 5

2024-05-31

Remote Sensing

Abstract:Indoor localization has long been a challenging task due to the complexity and dynamism of indoor environments. This paper proposes ULG-SLAM, a novel unsupervised learning and geometric-based visual SLAM algorithm for robot localizability estimation to improve the accuracy and robustness of visual SLAM. Firstly, a dynamic feature filtering based on unsupervised learning and moving consistency checks is developed to eliminate the features of dynamic objects. Secondly, an improved line feature extraction algorithm based on LSD is proposed to optimize the effect of geometric feature extraction. Thirdly, geometric features are used to optimize localizability estimation, and an adaptive weight model and attention mechanism are built using the method of region delimitation and region growth. Finally, to verify the effectiveness and robustness of localizability estimation, multiple indoor experiments using the EuRoC dataset and TUM RGB-D dataset are conducted. Compared with ORBSLAM2, the experimental results demonstrate that absolute trajectory accuracy can be improved by 95% for equivalent processing speed in walking sequences. In fr3/walking_xyz and fr3/walking_half, ULG-SLAM tracks more trajectories than DS-SLAM, and the ATE RMSE is improved by 36% and 6%, respectively. Furthermore, the improvement in robot localizability over DynaSLAM is noteworthy, coming in at about 11% and 3%, respectively.

environmental sciences,geosciences, multidisciplinary,imaging science & photographic technology,remote sensing

What problem does this paper attempt to address?

This paper proposes a new visual SLAM (Simultaneous Localization And Mapping) algorithm named ULG-SLAM, which combines unsupervised learning and geometric features. It aims to improve the accuracy and robustness of robot localization in indoor environments.

The paper mainly addresses the following issues:
1. **Impact of Dynamic Objects**: In real environments containing moving objects (such as pedestrians or vehicles), the accuracy of traditional visual SLAM systems is difficult to guarantee. Therefore, the paper proposes a dynamic feature filtering method based on unsupervised learning and motion consistency check to eliminate the impact of dynamic objects.
2. **Geometric Feature Extraction**: To compensate for the shortcomings of using only point features in scenes lacking texture or with significant lighting changes, the paper proposes an improved line feature extraction algorithm. This algorithm is based on LSD (Line Segment Detector) and optimizes the accuracy and effectiveness of line feature extraction.
3. **Localization Estimation Based on Geometric Features**: The paper also proposes a localization estimation optimization method based on geometric features and constructs an adaptive weight model and attention mechanism to significantly enhance the accuracy and robustness of localization estimation.

In summary, the main goal of this research is to improve the performance of existing visual SLAM systems in complex indoor environments by introducing unsupervised learning and enhancing geometric feature processing.