Video/AIS Fusion-Empowered Intelligent Maritime Surveillance in Inland Waterways__Ocean Engineering
The constructed multi-view vessel detection dataset is published on: https://pan.baidu.com/s/1vfws54AN5WiCgITGOR_hoQ Pass Code: mipc
If the MVV Dataset is used please cite it as:
Jingxiang Qu, Ryan Wen Liu, Yu Guo, Yuxu Lu, Jianlong Su, Peizheng Li, "Improving maritime traffic surveillance in inland waterways using the robust fusion of AIS and visual data", Ocean Engineering, 2023.
In our new work, the anti-occlusion algorithm is updated and integrated into a more accurate version: if the vessel is matched with its AIS, the occluded location will be predicted with the AIS data, if no matching, the location will be predicted via historical visual trajectory (as this paper introduced). The fusion algorithm is also updated to a trajecory-based matching method, which is more accurate. The source code will be avialable in our new paper:
Guo, Y., Liu, R. W., Qu, J., Lu, Y., Zhu, F., & Lv, Y. (2023). Asynchronous Trajectory Matching-Based Multimodal Maritime Data Fusion for Vessel Traffic Surveillance in Inland Waterways. IEEE Transactions on Intelligent Transportation Systems.
For more dataset about vessel detection the simultaneous data about video and AIS data fusion, please see our newest work: https://github.com/gy65896/FVessel, where the source code will be also available after the paper publishment.
Thanks for the following of our work!!!