Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques
This paper presents a camera-based vessel-speed enforcement system based on two cameras. The proposed system detects and tracks vessels per camera view and employs a re-identification (re-ID) function for linking vessels between the two cameras based on multiple bounding-box images per vessel. Newly...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-07-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/21/14/4659 |
| _version_ | 1797526059020713984 |
|---|---|
| author | Matthijs H. Zwemer Herman G. J. Groot Rob Wijnhoven Egor Bondarev Peter H. N. de With |
| author_facet | Matthijs H. Zwemer Herman G. J. Groot Rob Wijnhoven Egor Bondarev Peter H. N. de With |
| author_sort | Matthijs H. Zwemer |
| collection | DOAJ |
| description | This paper presents a camera-based vessel-speed enforcement system based on two cameras. The proposed system detects and tracks vessels per camera view and employs a re-identification (re-ID) function for linking vessels between the two cameras based on multiple bounding-box images per vessel. Newly detected vessels in one camera (query) are compared to the gallery set of all vessels detected by the other camera. To train and evaluate the proposed detection and re-ID system, a new Vessel-reID dataset is introduced. This extensive dataset has captured a total of 2474 different vessels covered in multiple images, resulting in a total of 136,888 vessel bounding-box images. Multiple CNN detector architectures are evaluated in-depth. The SSD512 detector performs best with respect to its speed (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>85.0</mn><mo>%</mo></mrow></semantics></math></inline-formula> Recall@95Precision at 20.1 frames per second). For the re-ID of vessels, a large portion of the total trajectory can be covered by the successful detections of the SSD model. The re-ID experiments start with a baseline single-image evaluation obtaining a score of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>55.9</mn><mo>%</mo></mrow></semantics></math></inline-formula> Rank-1 (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>49.7</mn><mo>%</mo></mrow></semantics></math></inline-formula> mAP) for the existing TriNet network, while the available MGN model obtains <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>68.9</mn><mo>%</mo></mrow></semantics></math></inline-formula> Rank-1 (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>62.6</mn><mo>%</mo></mrow></semantics></math></inline-formula> mAP). The performance significantly increases with <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>5.6</mn><mo>%</mo></mrow></semantics></math></inline-formula> Rank-1 (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>5.7</mn><mo>%</mo></mrow></semantics></math></inline-formula> mAP) for MGN by applying matching with multiple images from a single vessel. When emphasizing more fine details by selecting only the largest bounding-box images, another <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>2.0</mn><mo>%</mo></mrow></semantics></math></inline-formula> Rank-1 (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>1.4</mn><mo>%</mo></mrow></semantics></math></inline-formula> mAP) is added. Application-specific optimizations such as travel-time selection and applying a cross-camera matching constraint further enhance the results, leading to a final <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>88.9</mn><mo>%</mo></mrow></semantics></math></inline-formula> Rank-1 and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>83.5</mn><mo>%</mo></mrow></semantics></math></inline-formula> mAP performance. |
| first_indexed | 2024-03-10T09:24:45Z |
| format | Article |
| id | doaj.art-73ecae4d2f8e46b29a87605baa432469 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-03-10T09:24:45Z |
| publishDate | 2021-07-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-73ecae4d2f8e46b29a87605baa4324692023-11-22T04:54:08ZengMDPI AGSensors1424-82202021-07-012114465910.3390/s21144659Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques Matthijs H. Zwemer0Herman G. J. Groot1Rob Wijnhoven2Egor Bondarev3Peter H. N. de With4Department of Electrical Engineering, Eindhoven University of Technology, 5600 MB Eindhoven, The NetherlandsDepartment of Electrical Engineering, Eindhoven University of Technology, 5600 MB Eindhoven, The NetherlandsViNotion B.V., 5641 JA Eindhoven, The NetherlandsDepartment of Electrical Engineering, Eindhoven University of Technology, 5600 MB Eindhoven, The NetherlandsDepartment of Electrical Engineering, Eindhoven University of Technology, 5600 MB Eindhoven, The NetherlandsThis paper presents a camera-based vessel-speed enforcement system based on two cameras. The proposed system detects and tracks vessels per camera view and employs a re-identification (re-ID) function for linking vessels between the two cameras based on multiple bounding-box images per vessel. Newly detected vessels in one camera (query) are compared to the gallery set of all vessels detected by the other camera. To train and evaluate the proposed detection and re-ID system, a new Vessel-reID dataset is introduced. This extensive dataset has captured a total of 2474 different vessels covered in multiple images, resulting in a total of 136,888 vessel bounding-box images. Multiple CNN detector architectures are evaluated in-depth. The SSD512 detector performs best with respect to its speed (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>85.0</mn><mo>%</mo></mrow></semantics></math></inline-formula> Recall@95Precision at 20.1 frames per second). For the re-ID of vessels, a large portion of the total trajectory can be covered by the successful detections of the SSD model. The re-ID experiments start with a baseline single-image evaluation obtaining a score of <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>55.9</mn><mo>%</mo></mrow></semantics></math></inline-formula> Rank-1 (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>49.7</mn><mo>%</mo></mrow></semantics></math></inline-formula> mAP) for the existing TriNet network, while the available MGN model obtains <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>68.9</mn><mo>%</mo></mrow></semantics></math></inline-formula> Rank-1 (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>62.6</mn><mo>%</mo></mrow></semantics></math></inline-formula> mAP). The performance significantly increases with <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>5.6</mn><mo>%</mo></mrow></semantics></math></inline-formula> Rank-1 (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>5.7</mn><mo>%</mo></mrow></semantics></math></inline-formula> mAP) for MGN by applying matching with multiple images from a single vessel. When emphasizing more fine details by selecting only the largest bounding-box images, another <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>2.0</mn><mo>%</mo></mrow></semantics></math></inline-formula> Rank-1 (<inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>1.4</mn><mo>%</mo></mrow></semantics></math></inline-formula> mAP) is added. Application-specific optimizations such as travel-time selection and applying a cross-camera matching constraint further enhance the results, leading to a final <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>88.9</mn><mo>%</mo></mrow></semantics></math></inline-formula> Rank-1 and <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><mrow><mn>83.5</mn><mo>%</mo></mrow></semantics></math></inline-formula> mAP performance.https://www.mdpi.com/1424-8220/21/14/4659computer vision applicationvideo surveillancemaritime traffic managementvessel detectionvessel re-identification |
| spellingShingle | Matthijs H. Zwemer Herman G. J. Groot Rob Wijnhoven Egor Bondarev Peter H. N. de With Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques Sensors computer vision application video surveillance maritime traffic management vessel detection vessel re-identification |
| title | Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques |
| title_full | Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques |
| title_fullStr | Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques |
| title_full_unstemmed | Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques |
| title_short | Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques |
| title_sort | multi camera vessel speed enforcement by enhancing detection and re identification techniques |
| topic | computer vision application video surveillance maritime traffic management vessel detection vessel re-identification |
| url | https://www.mdpi.com/1424-8220/21/14/4659 |
| work_keys_str_mv | AT matthijshzwemer multicameravesselspeedenforcementbyenhancingdetectionandreidentificationtechniques AT hermangjgroot multicameravesselspeedenforcementbyenhancingdetectionandreidentificationtechniques AT robwijnhoven multicameravesselspeedenforcementbyenhancingdetectionandreidentificationtechniques AT egorbondarev multicameravesselspeedenforcementbyenhancingdetectionandreidentificationtechniques AT peterhndewith multicameravesselspeedenforcementbyenhancingdetectionandreidentificationtechniques |