Improving RGB-D SLAM accuracy in dynamic environments based on semantic and geometric constraints

Wang, Xiqi; Zheng, Shunyi; Lin, Xiao; Zhu, Fengbo

doi:10.1016/j.measurement.2023.113084

Cited by 13 publications

(4 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Solving the fundamental matrix requires constant iterations by randomly sampling 8 points, and the method easily fails when the number of dynamic feature points is large. Wang et al [27] proposed a geometric constraint model based on the correlation of feature point locations, which can effectively remove dynamically inconsistent feature points. However, the computation of correlation geometric constraints is relatively complex and can lead to computational delays in scenarios that require real-time processing of large amounts of data.…”

Section: Related Workmentioning

confidence: 99%

Real-time visual SLAM based YOLO-Fastest for dynamic scenes

Gong,

Sun,

Zou

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

Within the realm of autonomous robotic navigation, Simultaneous Localization and Mapping (SLAM) serves as a critical perception technology, drawing heightened attention in contemporary research. The traditional SLAM systems perform well in static environments, but in the real physical world, dynamic objects can destroy the static geometric constraints of the SLAM system, further limiting its practical application in the real world. In this paper, a robust dynamic RGB-D SLAM system is proposed to expand the number of static points in the scene by combining with YOLO-Fastest to ensure the effectiveness of the geometric constraints model construction, and then based on that, a new thresholding model is designed to differentiate the dynamic features in the objection bounding box, which takes advantage of the double polyline constraints and the residuals after reprojection to filter the dynamic feature points. In addition, two Gaussian models are constructed to segment the moving objects in the bounding box in the depth image to achieve the effect similar to the instance segmentation under the premise of ensuring the computational speed. In this paper, experiments are conducted on dynamic sequences provided by the TUM dataset to evaluate the performance of the proposed method, and the results show that the RMSE (Root mean squared error) metric of the absolute trajectory error of the algorithm of this paper has at least 80% improvement compared to ORB-SLAM2. Higher robustness in dynamic environments with both high and low dynamic sequences compared to DS-SLAM and Dynaslam, and can effectively provide intelligent localization and navigation for mobile robots.

show abstract

Section: Related Workmentioning

confidence: 99%

Real-time visual SLAM based YOLO-Fastest for dynamic scenes

Gong,

Sun,

Zou

et al. 2024

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…Mask-Fusion [27] employs geometric segmentation to generate more accurate object boundaries, thereby overcoming the imperfection of mask boundaries. SG-SLAM [28] is a dynamic RGB-D SLAM system based on Yolact and geometric constraints. It significantly enhances positioning accuracy; however, it relies on depth maps, and accuracy issues with Yolact make it unsuitable for outdoor environments.…”

Section: Dynamic Slammentioning

confidence: 99%

MOLO-SLAM: A Semantic SLAM for Accurate Removal of Dynamic Objects in Agricultural Environments

Lv,

Yao,

Guo

et al. 2024

Agriculture

View full text Add to dashboard Cite

Visual simultaneous localization and mapping (VSLAM) is a foundational technology that enables robots to achieve fully autonomous locomotion, exploration, inspection, and more within complex environments. Its applicability also extends significantly to agricultural settings. While numerous impressive VSLAM systems have emerged, a majority of them rely on static world assumptions. This reliance constrains their use in real dynamic scenarios and leads to increased instability when applied to agricultural contexts. To address the problem of detecting and eliminating slow dynamic objects in outdoor forest and tea garden agricultural scenarios, this paper presents a dynamic VSLAM innovation called MOLO-SLAM (mask ORB label optimization SLAM). MOLO-SLAM merges the ORBSLAM2 framework with the Mask-RCNN instance segmentation network, utilizing masks and bounding boxes to enhance the accuracy and cleanliness of 3D point clouds. Additionally, we used the BundleFusion reconstruction algorithm for 3D mesh model reconstruction. By comparing our algorithm with various dynamic VSLAM algorithms on the TUM and KITTI datasets, the results demonstrate significant improvements, with enhancements of up to 97.72%, 98.51%, and 28.07% relative to the original ORBSLAM2 on the three datasets. This showcases the outstanding advantages of our algorithm.

show abstract

“…Some scholars at home and abroad have conducted a lot of research. Wang Weiming et al proposed a large-space dynamic angle measurement method based on machine vision and servo control, which expanded the measuring range of a large space angle to 18.788 m [3]; Gan Yu et al proposed a dynamic frequency scanning interferometry method to separate and compensate the phase errors in measurement [4]; Esward applied modeling and simulation to dynamic measurement technology and found that it was helpful to improve the understanding of dynamic measurement tasks [5]; Zhang Qingsong et al used a 2D laser displacement sensor to achieve dynamic non-contact measurement of wheelset geometric parameters such as flange height and wheel diameter [6]; Wang Xiqi et al proposed a missing detection compensation algorithm based on a constant velocity model and a region growing algorithm to improve the semantic segmentation accuracy, which eliminated the impact of dynamic objects on visual SLAM trajectory accuracy [7]; Zhang Hongwen et al proposed a set of working time series for aerial surveying and mapping cameras and a new high-precision calibration method, the effectiveness of which was verified by simulation results and experimental data [8]; Li Lin et al applied embedded multi-scale deep learning to the dynamic performance measurement system of radio frequency identification, which can improve the reading distance of multiple tasks from the physical structure and collision resistance of the system [9]; Li Li analyzed and modeled the dynamic positioning errors of the CNC machine tool guide system, and proved that there is a regular variation between the dynamic positioning error of XY workbench and the movement velocity [10]; Li Guofang et al studied the automatic tool-changing system of a machining center and the dynamic force measurement system of the drag link mechanism to identify the time start and offset of each force value time in the dynamic force record [11]; Yang Juqing et al applied a fuzzy adaptive PID tracking algorithm to single-axis laser tracking and rapid prototype tracking experiments, which made the angular acceleration of dynamic tracking exceed 200 • /s 2 [12]. As the most important large-scale dynamic measuring instrument, a laser tracker must have a good dynamic tracking ability and accurate dynamic measurement ability [13].…”

Section: Introductionmentioning

confidence: 99%

Research on Radial Double Velocity Measurement Method of Laser Tracker

Lv,

Hu,

et al. 2024

Processes

View full text Add to dashboard Cite

For the dynamic problem that the low-speed sliding table is unable to meet the radial measuring speed of the laser tracker, this paper takes the sliding table of the indoor large-length standard device as the moving object to double the measuring distance by adding a pyramid prism, thereby doubling the radial speed of the laser tracker. In this paper, the measurement data are analyzed through equal interval measurement experiments, equal sampling frequency experiments and repeatability measurement experiments using a pyramid prism to obtain the following conclusions, respectively: Firstly, the stability of the actual interval of the laser tracker is optimal when the rated speed of the sliding table is 50 mm/s. When the pyramid prism is not used, the minimum standard deviation obtained by the laser tracker at a sampling interval of 5 mm is 0.0158 mm. Secondly, during the equal sampling frequency measurement, the stability of the laser interferometer is better than that of the laser tracker when the pyramid prism is not used. With two instruments at a sampling frequency of 10 Hz, the standard deviations of the local velocity of the laser interferometer are 0.1466 mm/s, 0.0693 mm/s and 0.1106 mm/s, respectively. The standard deviations of the laser tracker are 0.1582 mm/s, 0.1033 mm/s and 0.2008 mm/s, respectively. The same conclusion is obtained at a sampling frequency of 20 Hz. The stability of the laser tracker is better than that of the laser interferometer when the pyramid prism is used. Thirdly, the stability and reliability of the local velocity of the laser tracker are better than that of the laser interferometer. For example, at 10 Hz, the standard deviations of the local velocity of the laser tracker are 0.2745 mm/s and 0.2097 mm/s, respectively, and the repeatability of the local velocity is 0.9140 mm/s. The standard deviations of the local velocity of the interferometer are 0.6141 mm/s and 0.6368 mm/s, respectively, and the repeatability of the local velocity is 1.4886 mm/s. And the local velocity of the two measuring instruments is more reliable and stable at a low frequency (10 Hz). This experimental scheme provides a way to measure the high speed of a laser tracker using a sliding table at low speed.

show abstract

Improving RGB-D SLAM accuracy in dynamic environments based on semantic and geometric constraints

Cited by 13 publications

References 24 publications

Real-time visual SLAM based YOLO-Fastest for dynamic scenes

Real-time visual SLAM based YOLO-Fastest for dynamic scenes

MOLO-SLAM: A Semantic SLAM for Accurate Removal of Dynamic Objects in Agricultural Environments

Research on Radial Double Velocity Measurement Method of Laser Tracker

Contact Info

Product

Resources

About