摘要
This paper proposes a novel RGB-D SLAM algorithm which can recognize the features on static objects and estimate the ego-motion of RGB-D camera in dynamic scenes. Besides, a static score is established for each feature, which is used to indicate the static level of the feature. The static score consists of static density and static count. Static density represents the proportion of features of static objects around each feature, and static count represents the number of times each feature is recognized as a static feature. In the experimental section, we compare the proposed method with ORB-SLAM2 through the TUM RGB-D dataset. The results show that in dynamic scenes, the proposed method effectively improves the performance of the ego-motion of RGB-D camera.
学科与关键词
- Dynamic RGB-D SLAM Based on Static Probability and Observation NumberLiu, Yu; Wu, Yilin*; Pan, WenzhaoIEEE Transactions on Instrumentation and Measurement, 2021, 70: 8503411.被引 5
- Robust RGB: D-SLAM in highly dynamic environments based on probability observations and clustering optimizationLiu, Hailin; Tian, Liangfang*; Du, Qiliang; Xu, WenjieMeasurement Science and Technology, 2024, 35(3): 035405.
- Real-time motion removal based on point correlations for RGB-D SLAM in indoor dynamic environmentsWang, Kesai; Yao, Xifan*; Ma, Nanfeng; Jing, XuanNEURAL COMPUTING & APPLICATIONS, 2023, 35(12): 8707-8722.被引 1
