Extraction And Analysis Of Fishing Intensity Of Stow Net Fishing Vessels In The East Sea And Yellow Sea Based On Vms

Vessel Monitoring System(VMS)is an integrated application System for fishing vessels,which integrates fishing Vessel positioning,network communication,geographic information,data management,electronic information display and other technologies.Based on the Beidou Navigation System with independent intellectual property rights,China has built the Beidou fishing vessel monitoring and management service system.At present,Beidou VMS provides the information of longitude and latitude,dispatch time,speed and heading of fishing vessels with a time resolution of 3min and a spatial resolution of 10m,and the high-precision ship position information,which can be used for in-depth research in the aspects of fishing boat track identification,fishing status identification and operation area monitoring.In this study,Beidou VMS data of fishing vessels in Zhejiang Province in 2017 and provinces(municipalitie)along the East sea and yellow sea in 2018 were used.Firstly,the Beidou VMS ship position data were divided according to operating voyages,and the effective voyages were screened out and the operating characteristic track of the fishing vessel was drawn.Then,the characteristic tracks of the stow net and other fishing vessels are manually screened out,and the training set,verification set and test set are used to train the characteristic track recognition model.The complete position information of several trawler fishing vessels was selected randomly,and the fishing status of each position was manually marked.The classification model of fishing status was established by threshold screening,deep neural network and DBSCAN density clustering algorithm.Using this method on VMS data,information such as the number of nets,location and fishing duration of a trawler during a voyage can be obtained.Visualize the spatial distribution of fishing intensity of net fishing vessels in Zhejiang Sea in 2017 and East sea and yellow sea in 2018.The results are as follows:1)Stow net belongs to the fixed filter type net,which relies on the tide to extend the net body and filter the fish,so as to achieve the fishing purpose,and it is strongly dependent on the tide.The fishing vessels usually arrive at the target fishing ground before the beginning of the high tide(the first or 15th day of the lunar month)and return home after the end of fishing after the beginning of the low tide(the fifth or 20th day of the lunar month).The single-voyage concentrated operation area of the stow net is small,with a concentrated operation area of 30～100 km~2.This study used Beidou VMS ship position data of fishing vessels in Zhejiang Province in 2017 and fishing boats in six provinces(municipalitie)along the East sea and yellow sea in 2018 to pick up 733 and 4794 fishing trips of net fishing vessels,respectively.2)The fishing status of stow net fishing vessels can be divided into four states:navigation,net arrange/fishing,retrieve net and anchoring.The position data and characteristics of the vessels in different states are obviously different,and the arrange position of each net can be determined through the distribution of the vessel position state and the operation mode of the stow net fishing vessels.Using BP neural network to identify the fishing state has a high accuracy,but the accuracy of net location is low.The thresholding screening and DBSCAN density clustering methods were used to classify the fishing state with slightly lower accuracy,but the net position was judged more accurately.Finally,threshold screening,deep neural network and DBSCAN density clustering were used to form the ship position status discrimination model.The accuracy of ship position status judgment was 94.74%,and the accuracy of network position judgment was 93%.3)Determined by the state of fishing vessels fishing net position coordinates and calculate the fishing duration,calculate the longitude and latitude step length is 0.1°×0.1°accumulated fishing time within the grid,and divided by the geographic area of each grid,a 0.1°×0.1°grid average per square kilometer long fishing,as fishing intensity distribution,the unit(h/km~2).In 2017,the areas with net fishing intensity less than 10h/km~2 accounted for 51.44%of the total fishing area.The areas with fishing intensity greater than 10 h/km~2 and less than 20 h/km~2 accounted for 31.51%of the total fishing area.The areas with fishing intensity greater than 20 h/km~2 accounted for 17.03%of the total fishing area.In 2018,the area with fishing intensity less than 10 h/km~2 accounted for 69.71%of the total fishing area in the East sea and yellow sea.The areas with fishing intensity greater than 10 h/km~2 and less than 20 h/km~2 accounted for 13.31%of the total area.The areas with fishing intensity greater than 20 h/km~2 and less than 30 h/km~2accounted for 6.68%of the total area.The area with fishing intensity greater than 30h/km~2 accounted for 10.30%of the total area。4)The selectivity of stow net is poor,and the catch quantity of larval organisms with high economic value is large.Overfishing is not conducive to the recovery and sustainable utilization of fishery resources.At present,6 provinces(municipalitie)along the East Yellow Sea have imposed a 4.5-month compulsory fishing ban on the operation of fishing nets.In the first half of the year,the fishing intensity of net fishing vessels was relatively low and the distribution was relatively dispersed,and the fishing intensity was the lowest in February.The fishing intensity was relatively high in the second half of the year,and the areas with high fishing intensity were concentrated in October and November.The innovation points of this study are as follows:1)The Beidou VMS position data is used to extract the operating voyages of fishing vessels.The image recognition method is used to establish the track type recognition model and classify the operating voyages of net fishing vessels and other types of fishing vessels,so as to provide reference for the identification methods of fishing types of fishing vessels in the future.2)Threshold screening,deep neural network and DBSCAN density clustering algorithm were used to establish the fishing status discrimination model of the trawler,to judge the position status of the trawler in the operation voyage,and to determine the position of each net in the voyage,so as to fill the gap in the fishing status identification field of the stow net in China.3)Calculate the fishing duration of each net according to the ship position status,extract the distribution of net fishing intensity in the East sea and yellow sea,and discuss with environmental factors and relevant policies,so as to provide a new method for monitoring net fishing area and a new idea for the operation management of net fishing vessels.