Real-Time Closed-Loop Detection Method of vSLAM Based on a Dynamic Siamese Network
As visual simultaneous localization and mapping (vSLAM) is easy disturbed by the changes of camera viewpoint and scene appearance when building a globally consistent map, the robustness and real-time performance of key frame image selections cannot meet the requirements. To solve this problem, a rea...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-11-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/21/22/7612 |
| _version_ | 1797508457119612928 |
|---|---|
| author | Quande Yuan Zhenming Zhang Yuzhen Pi Lei Kou Fangfang Zhang |
| author_facet | Quande Yuan Zhenming Zhang Yuzhen Pi Lei Kou Fangfang Zhang |
| author_sort | Quande Yuan |
| collection | DOAJ |
| description | As visual simultaneous localization and mapping (vSLAM) is easy disturbed by the changes of camera viewpoint and scene appearance when building a globally consistent map, the robustness and real-time performance of key frame image selections cannot meet the requirements. To solve this problem, a real-time closed-loop detection method based on a dynamic Siamese networks is proposed in this paper. First, a dynamic Siamese network-based fast conversion learning model is constructed to handle the impact of external changes on key frame judgments, and an elementwise convergence strategy is adopted to ensure the accurate positioning of key frames in the closed-loop judgment process. Second, a joint training strategy is designed to ensure the model parameters can be learned offline in parallel from tagged video sequences, which can effectively improve the speed of closed-loop detection. Finally, the proposed method is applied experimentally to three typical closed-loop detection scenario datasets and the experimental results demonstrate the effectiveness and robustness of the proposed method under the interference of complex scenes. |
| first_indexed | 2024-03-10T05:05:13Z |
| format | Article |
| id | doaj.art-a33d607bb1ed425686929191717e6168 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T05:05:13Z |
| publishDate | 2021-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-a33d607bb1ed425686929191717e61682023-11-23T01:26:42ZengMDPI AGSensors1424-82202021-11-012122761210.3390/s21227612Real-Time Closed-Loop Detection Method of vSLAM Based on a Dynamic Siamese NetworkQuande Yuan0Zhenming Zhang1Yuzhen Pi2Lei Kou3Fangfang Zhang4School of Computer Technology and Engineering, Changchun Institute of Technology, Changchun 130012, ChinaSchool of Electrical Engineering, Northeast Electric Power University, Jilin 132011, ChinaNational Local Joint Engineering Research Center for Smart Distribution, Grid Measurement and Control with Safety Operation Technology, Changchun Institute of Technology, Changchun 130012, ChinaInstitute of Oceanographic Instrumentation, Qilu University of Technology (Shandong Academy of Sciences), Qingdao 266075, ChinaSchool of Electrical Engineering and Automation, Qilu University of Technology (Shandong Academy of Sciences), Jinan 250353, ChinaAs visual simultaneous localization and mapping (vSLAM) is easy disturbed by the changes of camera viewpoint and scene appearance when building a globally consistent map, the robustness and real-time performance of key frame image selections cannot meet the requirements. To solve this problem, a real-time closed-loop detection method based on a dynamic Siamese networks is proposed in this paper. First, a dynamic Siamese network-based fast conversion learning model is constructed to handle the impact of external changes on key frame judgments, and an elementwise convergence strategy is adopted to ensure the accurate positioning of key frames in the closed-loop judgment process. Second, a joint training strategy is designed to ensure the model parameters can be learned offline in parallel from tagged video sequences, which can effectively improve the speed of closed-loop detection. Finally, the proposed method is applied experimentally to three typical closed-loop detection scenario datasets and the experimental results demonstrate the effectiveness and robustness of the proposed method under the interference of complex scenes.https://www.mdpi.com/1424-8220/21/22/7612simultaneous localization and mappingclosed-loop detectionSiamese networkdeep learningelementwise integration strategy |
| spellingShingle | Quande Yuan Zhenming Zhang Yuzhen Pi Lei Kou Fangfang Zhang Real-Time Closed-Loop Detection Method of vSLAM Based on a Dynamic Siamese Network Sensors simultaneous localization and mapping closed-loop detection Siamese network deep learning elementwise integration strategy |
| title | Real-Time Closed-Loop Detection Method of vSLAM Based on a Dynamic Siamese Network |
| title_full | Real-Time Closed-Loop Detection Method of vSLAM Based on a Dynamic Siamese Network |
| title_fullStr | Real-Time Closed-Loop Detection Method of vSLAM Based on a Dynamic Siamese Network |
| title_full_unstemmed | Real-Time Closed-Loop Detection Method of vSLAM Based on a Dynamic Siamese Network |
| title_short | Real-Time Closed-Loop Detection Method of vSLAM Based on a Dynamic Siamese Network |
| title_sort | real time closed loop detection method of vslam based on a dynamic siamese network |
| topic | simultaneous localization and mapping closed-loop detection Siamese network deep learning elementwise integration strategy |
| url | https://www.mdpi.com/1424-8220/21/22/7612 |
| work_keys_str_mv | AT quandeyuan realtimeclosedloopdetectionmethodofvslambasedonadynamicsiamesenetwork AT zhenmingzhang realtimeclosedloopdetectionmethodofvslambasedonadynamicsiamesenetwork AT yuzhenpi realtimeclosedloopdetectionmethodofvslambasedonadynamicsiamesenetwork AT leikou realtimeclosedloopdetectionmethodofvslambasedonadynamicsiamesenetwork AT fangfangzhang realtimeclosedloopdetectionmethodofvslambasedonadynamicsiamesenetwork |