TOKYO: Mitsui O.S.K. Lines, Ltd. announced the joint development, along
with SenseTime Japan Ltd., of a new vessel image recognition and
recording system, and the system’s installation for demonstration
testing aboard the cruise ship “Nippon Maru” operated by Mitsui O.S.K.
Passenger Line, Ltd. The system’s graphic recognition engine was
developed by incorporating MOL’s knowledge and applying artificial
intelligence (AI) deep learning technology, allowing it to recognize
surrounding vessels. The system recognizes and automatically records
vessels with high accuracy using a terminal equipped with a graphic
processing unit (GPU) and ultrahigh-resolution cameras.
The image recognition technology, which can be used even at night or
during other periods of poor visibility, can detect small vessels
(International treaty and domestic law require vessels to be equipped
with AIS, but the regulations do not apply to ocean-going freighters of
less than 300 gross tonnes and coastal vessels of less than 500 gross
tonnes) that are not recognizable by vessel automatic identification
systems (AIS). MOL views advanced image recognition as an underlying
technology needed for automated watch-keeping, as the company works to
realize the autonomous smart ships of the future.
Since this system automatically records image data, MOL plans to
carefully review the accumulated data and use it to further enhance the
image recognition engine’s analysis accuracy.
MOL positions the navigation support system featuring image
recognition, such as the AR voyage information display system , which is
already in practical use, as part of its challenge project, called the
“FOCUS EYE” series. The company continues to develop and modify the
system, which supports crewmembers’ watch-keeping while underway, to
enhance safe operation and speed up the automation of watch-keeping,
with the long-term goal of realizing autonomous smart ships, said a MOL
release.