RGB-D Visual SLAM Based on Yolov4-Tiny in Indoor Dynamic Environment

Chang, Zhanyuan; Wu, Honglin; Sun, Yunlong; Li, Chuanjiang

doi:10.3390/mi13020230

Cited by 15 publications

(4 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Root Mean Square Error (RMSE) is then employed to compute the overall value of this error. ATE and RPE are two parameters that reflect the robustness and stability of a visual SLAM system [9]. A lower RMSE value calculated from ATE and RPE indicates a better fitting performance.…”

Section: Visual Slam Front-end Feature Extraction Effect After Integr...mentioning

confidence: 99%

Optimization method to improve visual SLAM in dynamic environment

Liu,

Ishii

2024

Proceedings of International Conference on Artificial Life and Robotics

View full text Add to dashboard Cite

Aiming at the problem that the robustness of the classic visual SLAM system is greatly affected by dynamic target feature points in the environment, a method is proposed to use a target detection algorithm to identify and eliminate dynamic target feature points. First, use the target detection algorithm YOLOv5 to identify the collected environmental images, and select the surrounding environment. Objects identified as dynamic targets in the environment, and then the target detection results are integrated into the feature extraction of the visual SLAM frontend, the feature points belonging to the dynamic target part of the extracted image feature points are removed, and the remaining static feature points are used to map Construction and positioning, and finally testing on the TUM data set. The results show that after using the target detection algorithm to eliminate dynamic feature points, the root mean square error of the absolute trajectory error of the visual SLAM system in highly dynamic scenes is reduced by 97.89%, effectively improving the positioning accuracy and robustness of the system.

show abstract

Section: Visual Slam Front-end Feature Extraction Effect After Integr...mentioning

confidence: 99%

Optimization method to improve visual SLAM in dynamic environment

Liu,

Ishii

2024

Proceedings of International Conference on Artificial Life and Robotics

View full text Add to dashboard Cite

show abstract

“…Since significant progress has been made in visual SLAM methods, Davison et al [3]introduced the first fully visual SLAM method based on monocular cameras, MonoSLAM is a visual SLAM technique for realtime applications based on the feature method, but the computational complexity is proportional to the number of feature points, so it can only be used in small scenes; Klein and Murray [4] proposed a SLAM system called PTAM, which separates tracking and mapping, and the method effectively reduces the computational cost of monocular SLAM; Mur-Artal et al [5,6] first proposed the open source monocular ORB-SLAM system, which was investigated in the algorithmic framework of [4]; Mur-Artal et al [6] then proposed the ORB-SLAM2 system, supporting binocular and RGBD cameras based on ORB-SLAM, which was a significant advancement in visual SLAM and one of the most widely used contemporary methods due to its robustness [7]. Combining motion-compensated image difference and maximum a posteriori estimation, Sun et al [8] proposed a method to roughly detect moving objects.…”

Section: Slam (Simultaneous Localization and Mappingmentioning

confidence: 99%

DDETR-SLAM: A Transformer-Based Approach to Pose Optimization in Dynamic Environments

Liu

Wang

et al. 2023

Preprint

View full text Add to dashboard Cite

Simultaneous Localization and Mapping (SLAM) is a critical technology for accurate robot localization and path planning. It has been an important area of research to improve localization accuracy. In this paper, we propose a Transformer-based visual semantic SLAM algorithm (DDETR-SLAM) to address the shortcomings of traditional visual SLAM frameworks, such as large localization errors in dynamic scenes and “ghosting” in 3D mapping. First, by incorporating the Deformable DETR (DEtection TRansformer) network as an object detection thread, the pose estimation accuracy of the system is improved compared to ORB-SLAM2. Furthermore, a dynamic feature point culling algorithm that combines the semantic information is designed to eliminate outlier points generated by dynamic objects, thereby improving the accuracy and robustness of SLAM localization and mapping. Experiments are conducted on the public TUM datasets to verify the localization accuracy, computational efficiency, and readability of the point cloud map of DDETR-SLAM. The results show that in highly dynamic environments, the ATE (Absolute Trajectory Error), translation error, and rotation error are reduced by 98.45%, 95.34%, and 92.67%, respectively, when compared to ORB-SLAM2. In most cases, our proposed system outperforms DS-SLAM, DynaSLAM, Detect-SLAM, RGB-D SLAM, and YOLOv5 + ORB-SLAM2, and our methodology improves location accuracy. The dense mapping also has better readability. The RPE (Relative Trajectory Error) is only 0.0076 m, and the ATE is only 0.0063 m.

show abstract

“…Typical methods for point feature matching can be summarized as descriptor-based and pixel-based methods [ 21 ], due to the difficulty of establishing sufficient dynamic data associations, the descriptor-based method can be challenging to apply to highly dynamic objects [ 22 ]. As such, many systems resort to using the optical flow method to address this issue [ 23 , 24 , 25 ], outliers are then removed from the essential matrix using the RANSAC method [ 26 ].…”

Section: Introductionmentioning

confidence: 99%

An Adaptive ORB-SLAM3 System for Outdoor Dynamic Environments

Zang

Zhang

2023

Sensors

View full text Add to dashboard Cite

Recent developments in robotics have heightened the need for visual SLAM. Dynamic objects are a major problem in visual SLAM which reduces the accuracy of localization due to the wrong epipolar geometry. This study set out to find a new method to address the low accuracy of visual SLAM in outdoor dynamic environments. We propose an adaptive feature point selection system for outdoor dynamic environments. Initially, we utilize YOLOv5s with the attention mechanism to obtain a priori dynamic objects in the scene. Then, feature points are selected using an adaptive feature point selector based on the number of a priori dynamic objects and the percentage of a priori dynamic objects occupied in the frame. Finally, dynamic regions are determined using a geometric method based on Lucas-Kanade optical flow and the RANSAC algorithm. We evaluate the accuracy of our system using the KITTI dataset, comparing it to various dynamic feature point selection strategies and DynaSLAM. Experiments show that our proposed system demonstrates a reduction in both absolute trajectory error and relative trajectory error, with a maximum reduction of 39% and 30%, respectively, compared to other systems.

show abstract

RGB-D Visual SLAM Based on Yolov4-Tiny in Indoor Dynamic Environment

Cited by 15 publications

References 35 publications

Optimization method to improve visual SLAM in dynamic environment

Optimization method to improve visual SLAM in dynamic environment

DDETR-SLAM: A Transformer-Based Approach to Pose Optimization in Dynamic Environments

An Adaptive ORB-SLAM3 System for Outdoor Dynamic Environments

Contact Info

Product

Resources

About