PLD-SLAM: A New RGB-D SLAM Method with Point and Line Features for Indoor Dynamic Scene

Chenyang Zhang,Teng Huang,Rongchun Zhang,Xuefeng Yi

DOI: https://doi.org/10.3390/ijgi10030163

2021-03-13

Abstract:RGB-D SLAM (Simultaneous Localization and Mapping) generally performs smoothly in a static environment. However, in dynamic scenes, dynamic features often cause wrong data associations, which degrade accuracy and robustness. To address this problem, in this paper, a new RGB-D dynamic SLAM method, PLD-SLAM, which is based on point and line features for dynamic scenes, is proposed. First, to avoid under-over segmentation caused by deep learning, PLD-SLAM combines deep learning for semantic information segmentation with the K-Means clustering algorithm considering depth information to detect the underlying dynamic features. Next, two consistency check strategies are utilized to check and filter out the dynamic features more reasonably. Then, to obtain a better practical performance, point and line features are utilized to calculate camera pose in the dynamic SLAM, which is also different from most published dynamic SLAM algorithms based merely on point features. The optimization model with point and line features is constructed and utilized to calculate the camera pose with higher accuracy. Third, enough experiments on the public TUM RGB-D dataset and the real-world scenes are conducted to verify the location accuracy and performance of PLD-SLAM. We compare our experimental results with several state-of-the-art dynamic SLAM methods in terms of average localization errors and the visual difference between the estimation trajectories and the ground-truth trajectories. Through the comprehensive comparisons with these dynamic SLAM schemes, it can be fully demonstrated that PLD-SLAM can achieve comparable or better performances in dynamic scenes. Moreover, the feasibility of camera pose estimation based on both point features and line features has been proven by the corresponding experiments from a comparison with our proposed PLD-SLAM only based on point features.

geography, physical,remote sensing,computer science, information systems

What problem does this paper attempt to address?

The paper proposes a new RGB-D Simultaneous Localization and Mapping (SLAM) method—PLD-SLAM, which is designed for indoor dynamic scenes and combines point features and line features. The paper aims to address the following issues:

1. **Limitations of traditional SLAM in dynamic environments**: Most existing SLAM systems perform well in static environments, but in dynamic scenes, the presence of dynamic objects leads to incorrect data association, thereby reducing the accuracy and robustness of localization and mapping.

2. **Feature detection and classification**: The paper proposes a new method to distinguish between static and dynamic features, avoiding the over-segmentation problem caused by deep learning, and retaining as many static features as possible to improve localization accuracy.

3. **Optimization using point and line features**: Most SLAM algorithms use only point features, while PLD-SLAM creatively combines point features and line features to construct an optimization model for more accurate camera pose estimation.

Specifically, the contributions of the paper include:
- Proposing a new RGB-D SLAM method (PLD-SLAM) that combines point features and line features, and uses the MobileNet model and K-Means clustering algorithm to detect potential dynamic feature points and lines.
- Further identifying true dynamic feature points through two consistency check strategies and retaining as many static features as possible.
- Constructing an optimization model using point and line features to achieve higher accuracy in camera pose estimation.
- Conducting extensive experiments on the public TUM RGB-D dataset to verify the localization accuracy and performance of PLD-SLAM in dynamic scenes, and comparing it with some advanced dynamic SLAM methods, demonstrating its superiority or comparable performance in dynamic scenes.