MSC Cruises has added a man overboard detection system to one of its cruise ships.
The system, claimed to be the first in the industry, will eventually be rolled out to all company cruise ships.
The video surveillance system was developed in collaboration with security technology companies, Bosch and Hewlett Packard Enterprise, and will allow for rapid intervention in the unlikely event a person or object falls overboard.
It is part of a larger MSC Cruises operation to further optimise security monitoring on board. After a successful pilot phase of extensive stress-testing and continuous software upgrades, the intelligent video capturing and analysis system is now in operation – on MSC Cruises’ ‘MSC Meraviglia’.
The system consists of a shield of intelligent optical and thermal video cameras, which provide non stop comprehensive surveillance alongside the relevant exterior parts of the ship.
All captured video images are streamed in real time to a central security room where the video stream is monitored together with all other inputs from the 1,200 HD CCTV cameras on board.
MSC Cruises’ security department and experts from Bosch and Hewlett Packard Enterprise have worked over 14 months to meet the security requirements and create the right surveillance setup.
While a key component in the surveillance infrastructure is to install highly resilient hardware (cameras, casing, and bases) that resists tough conditions at sea, a significant part of the process is to program the software so that it provides reliable and accurate data.
In order to further increase the accuracy of the captured video images, they are processed and analysed by two separate and independent image processing systems.
This double security system has allowed to significantly lower the error margin for false alerts – typically caused by natural movements, such as waves, reflection of sun or moon, or a bird triggering the alarm. These interferences usually pose significant challenges for current shipboard CCTV technology.
Through over 25,000 hours of video analysis, extensive software testing and continuous algorithmic updates, the system has now reached a confirmed accuracy level of 97%, the company claimed.