This project is a research and development project aiming at achieving new changes to the ocean and fishery industry as a new application of pattern recognition technology through collaboration of top researchers of marine science, fisheries science and information science.
“Fishing boat navigation” contributing to productivity improvement of sea surface fishery
We will support the efficiency of fishery activities by providing frequent and highly accurate sea status information rather than what has been provided in the past. We predict the fishing ground by analyzing the water lifting data and the seawater temperature distribution for each operation point, support the decision of the fishing ground determined by experience and intuition, and navigate to the fishing ground.
Technology supporting “fishing boat navigation”
Prediction of wide area oceanic condition based on ocean physical model It is a technology to predict the sea condition such as sea water temperature over a wide range by model calculation using a large computer.
Wide area prediction of the ocean is obtained by optimizing observation data that satellites, observation buoys, and observation vessels provide based on the ocean physical model. This calculation requires a large amount of computer resources, and the latest supercomputer is used.
OnSpot ocean condition forecast
Using data assimilation technology, predict the ocean condition around the fishing boat on demand.
Although prediction of the ocean condition based on the physical model of the sea is highly accurate, it was difficult to forecast the range that can be directly used for fishing activities. We have developed a technology to realize the prediction range of 10 km units in the finer range of 100 m by utilizing data from sensors mounted on vessels such as fishing boats. With this technology, it is now possible to provide more useful information for fishery activities, such as determining more detailed fishing ground decisions and netting timing.
Automatic creation of isotherm diagram
We will remove the clouds from the satellite water temperature image at high speed and create the isotherm diagram as quickly as possible.
In the sea fishery field, decisions on which seas to fish are important decision-making items affecting landing and fuel consumption. Isotherm diagrams provided by public institutions etc. are widely used for fishing ground decisions. By using the image reconstruction technique of the deep learning system, we greatly reduced the calculation cost and automated the preparation of the isotherm diagram itself than before. This greatly improved the frequency of renewal and the timing of renewal, making it possible to make a big contribution to improving the efficiency of fishery.
Estimated fishing veteran fisherman’s experience x Pattern analysis
At the fishing site, based on isotherm charts etc, veteran fishermen decide the sea area to be fished by experience and intuition. We developed a technique to imitate the skill of veteran fishermen and to estimate the fishery which is well captured by performing pattern recognition of catch data and seawater temperature distribution.