Identification of fishing type from VMS data based on artificial neural network

Unreasonable fishing ways lead to decay of marine fishery resources and destruction of marine ecological environment. In re- cent years, vessel monitoring system has been used for vessel safety supervision, fishery resources management, marine ecological envi- ronment protection, etc. The paper selects 78 vessels fishing in offshore China, including 15 flow gill net fishing boats, 39 trawlers and 24 flow stow net fishing boats and used BP neural network as model to identify fishing type by speed and azimuth from Beidou VMS data in 2014. Results show that the correct rates of identification based on speed were 93. 6% and 91% , both could classify fishing types well. For trawler and flow stow net fishing boats, the correct classification were both over 90% , but that of flow gill net fishing boat was only about 70% , which might be resulted from insufficient network training, or speed and lack of characteristic azimuth data.