What is it about?
the team proposes an improved deep learning YOLOv5s algorithm[1], which can greatly improve the detection accuracy of crew's unsafe behaviors and improve the detection efficiency of single frame[5], and integrated it with the front-end camera terminal to design and propose a ship distributed safety supervision system[4] based on front-end intelligence. It can effectively overcome the defects of traditional supervision, such as low efficiency, high cost and delayed data reception.
Featured Image
Photo by Venti Views on Unsplash
Why is it important?
At present, it is common to use video terminals to monitor the status of personnel on duty in different areas above and below the deck. However, there are still deficiencies of low degree of intelligence and high degree of data center. The details are as follows:When the central server fails, it is easy to cause the collapse of the whole ship supervision system; If the crew members have unsafe behaviors on duty (like smoking, playing mobile phones during duty, absence from work), the traditional centralized monitoring cannot actively identify the risk behaviors, resulting in increased risk uncertainty; It takes 24 hours to collect video surveillance information continuously, occupies human resources, and requires manual review and playback, resulting in low reliability and efficiency; Data transmission and processing are slow, and there is network delay. Based on the above deficiencies, the team proposes an improved deep learning YOLOv5s algorithm[1], which can greatly improve the detection accuracy of crew's unsafe behaviors and improve the detection efficiency of single frame[5], and integrated it with the front-end camera terminal to design and propose a ship distributed safety supervision system[4] based on front-end intelligence. It can effectively overcome the defects of traditional supervision, such as low efficiency, high cost and delayed data reception.
Perspectives
Writing this article was a great pleasure as I had learned much during that times.
Ziyun Luo
Shanghai Maritime University
Read the Original
This page is a summary of: Design and implementation of ship distributed safety supervision system based on front-end intelligence, December 2022, ACM (Association for Computing Machinery),
DOI: 10.1145/3584376.3584514.
You can read the full text:
Contributors
The following have contributed to this page
