M3C: Multimodel-and-Multicue-Based Tracking by Detection of Surrounding Vessels in Maritime Environment for USV

Qiao, Dalei; Liu, Guangzhong; Zhang, Jun; Zhang, Qiangyong; Wu, Gongxing; Feng, Dazeng

doi:10.3390/electronics8070723

Cited by 15 publications

(9 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The work of Qiauo et al [ 4 ] concentrates on unrestricted detection, tracking and re-ID of moving vessels, to recognize other vessels in their surroundings. They use a CNN-based detection model [ 5 ] and apply sensor fusion and a re-ID algorithm to achieve accurate vessel tracking in their surroundings.…”

Section: Related Workmentioning

confidence: 99%

Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques

Zwemer

Groot

Wijnhoven

et al. 2021

Sensors

View full text Add to dashboard Cite

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 (85.0% 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 55.9% Rank-1 (49.7% mAP) for the existing TriNet network, while the available MGN model obtains 68.9% Rank-1 (62.6% mAP). The performance significantly increases with 5.6% Rank-1 (5.7% 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 2.0% Rank-1 (1.4% 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 88.9% Rank-1 and 83.5% mAP performance.

show abstract

Section: Related Workmentioning

confidence: 99%

Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques

Zwemer

Groot

Wijnhoven

et al. 2021

Sensors

View full text Add to dashboard Cite

show abstract

“…Heyse et al developed a method of identifying marine vessels using multi-level descriptions and created a refined multi-level classifier based on deep features [22]. Qiao et al addressed ship re-identification for the long-term tracking of vessels at sea by considering each image as a set of visual cues [23]. Tian et al used a word bag model to recognize depth features [24] and demonstrated its use on a large image gallery.…”

Section: B Marine Vessel Detection and Identificationmentioning

confidence: 99%

Marine Vessel Re-Identification: A Large-Scale Dataset and Global-and-Local Fusion-Based Discriminative Feature Learning

et al. 2020

View full text Add to dashboard Cite

A marine vessel re-identification system has to determine whether or not different images represent the same vessel. Accurate vessel re-identification improves onshore closed-circuit television monitoring in a vessel traffic services system as well as onboard surveillance of surrounding vessels. However, because ships are rigid bodies and the marine environment is harsh, the accurate re-identification of vessels at sea can be very difficult. We describe a marine vessel-re-identification framework, Globaland-Local Fusion-based Multi-view Feature Learning (GLF-MVFL), which is based on a combination of global and fine-grained local features. GLF-MVFL combines cross-entropy loss with our newly-developed orientation-guided quintuplet loss. We exploit intrinsic features of marine vessels to optimize multi-view representation learning for re-identification. GLF-MVFL uses ResNet-50 as the backbone network to extract features for simultaneous quintuple input. It detects and discriminates between features and estimates viewpoints to form a comprehensive re-identification framework. We created an annotated large-scale vessel retrieval dataset, VesselID-539, which contains images from viewpoints similar to those of an autonomous surface vessel, to use in evaluating the performance of the model. Extensive experiments and analysis of the results obtained from using VesselID-539 demonstrate that our approach significantly increases the accuracy of vessel re-identification and is more effective and robust for images from different viewpoints than other approaches. INDEX TERMS Autonomous surface vessel (ASV), maritime surveillance, VesselID-539 dataset, multi views, vessel re-identification (V-ReID).

show abstract

“…32 If the influence of ship state and historical behavior are given, LSTM network will be an appropriate choice of historical sequence relevance that can remember the critical data and forget the less-important data. 33 Hence, the combination of CNN and LSTM can be taken as the self-learning framework between ship state and behavior prediction, 34 whose structure solves the key problems of uncertainty and adaptability.…”

Section: Deep Learning Network Structurementioning

confidence: 99%

Deep learning structure for collision avoidance planning of unmanned surface vessel

Zheng

2020

Proceedings of the Institution of Mechanical Engineers, Part M:

View full text Add to dashboard Cite

The uncertain dynamics and environmental impacts in collision avoidance planning are always the key issues of autonomous navigation for unmanned surface vessel, due to the strong interactive ability and learning ability of deep learning, the combination of CNN and LSTM provide relevance and memory characteristics, which can extract the potential influence features of collision avoidance and correlation of historical strategies, to gain the more general trend of collision avoidance. Meanwhile, to make sure the real-time problem, the heuristic planning based on improved Electromagnetism-like mechanism algorithm is utilized to seek the optimal strategy with the rolling iteration of CNN–LSTM structure, for obtaining a simple and effective on-line scheme of collision avoidance. The complete algorithm has been tested and validated under the different situation comparing with the deterministic algorithm, it is worth mentioning that the algorithm can potentially be developed to advance collision-free planning especially in multiple vessels situation.

show abstract

M3C: Multimodel-and-Multicue-Based Tracking by Detection of Surrounding Vessels in Maritime Environment for USV

Cited by 15 publications

References 48 publications

Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques

Multi-Camera Vessel-Speed Enforcement by Enhancing Detection and Re-Identification Techniques

Marine Vessel Re-Identification: A Large-Scale Dataset and Global-and-Local Fusion-Based Discriminative Feature Learning

Deep learning structure for collision avoidance planning of unmanned surface vessel

Contact Info

Product

Resources

About