A dense map optimization method based on common-view geometry

Jiang, Luomin; Wang, Congqing; Luo, Dehan

doi:10.1007/s11760-020-01846-6

Cited by 3 publications

(3 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Equation (15), F denotes the basis matrix for the transformation of coordinate O 1 to coordinate O 2 and E is the essence matrix. Therefore, the rotation matrix R and the translation vector t between the two coordinate systems can be calculated according to Equation (13).…”

Section: Principle Of Dynamic Feature Removal Methodsmentioning

confidence: 99%

“…There are two main approaches for dynamic objects in the environment. (1), using the geometric approach [13] for dynamic feature removal, and (2), using the semantic segmentation approach [14] for dynamic feature removal. Chenyang Zhang et al [15] proposed a non-semantic a priori dynamic feature proposed method based on particle filtering for dynamic feature removal in the environment.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Geometric Constraint-Based and Improved YOLOv5 Semantic SLAM for Dynamic Scenes

Zhang

2023

IJGI

View full text Add to dashboard Cite

When using deep learning networks for dynamic feature rejection in SLAM systems, problems such as a priori static object motion leading to disturbed build quality and accuracy and slow system runtime are prone to occur. In this paper, based on the ORB-SLAM2 system, we propose a method based on improved YOLOv5 networks combined with geometric constraint methods for SLAM map building in dynamic environments. First, this paper uses ShuffleNetV2 to lighten the YOLOv5 network, which increases the improved network’s operation speed without reducing the accuracy. At the same time, a pyramidal scene parsing network segmentation head is added to the head part of the YOLOv5 network to achieve semantic extraction in the environment, so that the improved YOLOv5 network has both target detection and semantic segmentation functions. In order to eliminate the objects with low dynamic features in the environment, this paper adopts the method of geometric constraints to extract and eliminate the dynamic features of the low dynamic objects. By combining the improved YOLOv5 network with the geometric constraint method, the robustness of the system is improved and the interference of dynamic targets in the construction of the SLAM system map is eliminated. The test results on the TUM dataset show that, when constructing a map in a dynamic environment, compared with the traditional ORB-SLAM2 algorithm, the accuracy of map construction in a dynamic environment is significantly improved. The absolute trajectory error is reduced by 97.7% compared with ORB-SLAM2, and the relative position error is reduced by 59.7% compared with ORB-SLAM2. Compared with DynaSLAM for dynamic scenes of the same type, the accuracy of map construction is slightly improved, but the maximum increase in keyframe processing time is 94.7%.

show abstract

Section: Principle Of Dynamic Feature Removal Methodsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Geometric Constraint-Based and Improved YOLOv5 Semantic SLAM for Dynamic Scenes

Zhang

2023

IJGI

View full text Add to dashboard Cite

show abstract

“…With the development of the global navigation satellite system (GNSS), GNSS time transfer is widely used in the elds of time scale maintenance, scienti c studies, and dense map optimization [1][2][3][4] with its advantages of global, all-day, being able to operate continuously, high precision, and exibility. In the future, it may also be combined with location-based service (LBS) for private needs [5,6], which can provide support for heterogeneous networks, indoor navigation, and in-vehicle Internet [7][8][9].…”

Section: Introductionmentioning

confidence: 99%

Modelling and Assessment of BDS-3 Real-time Precise Point Positioning Time Transfer Based on PPP-B2b Service

Tang

Lyu

Zeng

et al. 2022

Mathematical Problems in Engineering

View full text Add to dashboard Cite

In 2020, users in and around China had access to a new real-time precise point positioning (RTPPP) service with the BeiDou global navigation satellite system (BDS-3) B2b (PPP-B2b) signal. In this study, the quality of PPP-B2b products is first assessed and compared with the CNAV1 broadcast ephemeris. Then a mode of BDS-3 PPP time transfer with PPP-B2b (B2b-RTPPP) is developed and evaluated in static and kinematic modes. The results demonstrate that the discontinuity of orbit is improved by applying the PPP-B2b, and its root mean square errors (RMSEs) are 0.081 m, 0.165 m, and 0.107 m in the radial, along-track, and cross-track components, respectively. The standard deviation (STD) of the PPP-B2b clock offset is 0.08 ns, greatly better than that of the broadcast clock offset. For time transfer, the type A uncertainty of the B2b-RTPPP solution is approximately 0.1 ns in static mode, and approximately 0.3 ns in kinematic mode. The B1I/B3I B2b-RTPPP time transfer solution performs better than the B1C/B2a solution. For frequency stability, the modified Allan deviation (MDEV) of the B2b-RTPPP solution is comparable to that of the post-processing solution. The B1C/B2a combination has better short-term frequency stability than B1I/B3I, while the long-term frequency stability of B1C/B2a is worse. In addition, the contribution of the differential code bias (DCB) to B2b-RTPPP time transfer is also investigated. The type A uncertainty of the B2b-RTPPP solution without DCB correction is worse than that of the B2b-RTPPP solution with DCB correction. The B1C/B2a B2b-RTPPP solution without DCB correction has better frequency stability than the B1I/B3I B2b-RTPPP solution.

show abstract