Research Of Tuna Fish-tracking System Based On Sonar Image Data For Seine Fishing

A key technology in the tuna purse seine is to master underwater information withthe use of sonar. Sonar has changed from the search of fish into the main purposes formonitoring the status of the underwater seine net shape and fish movement. Tuna purseseine fishery has high rate of empty catch, and it’s hard to master the underwater fishmovements. It is particularly important if we can do the real-time tracking of fishmovement, predict fish speed and take measures to prevent fish escape.This paperpresents a tuna tracking system based on sonar acoustic data, which is composed of asonar, data acquisition system and fish tracking algorithm. Now the proposed systemhas been installed in a tuna fishing vessel for recording real data.The purpose of this study is to investigate the application of tracking system basedon the sonar images of tuna, hope to achieve the tracking state of tuna fish movement,improve the success rate of seine. The hardware integration of the tuna tracking systemis composed with the sonar, data acquisition system and recording system, the softwareintegration is developed for the acoustic image processing methods such as thecomprehensive application of IMM-JPDA tracking algorithm, which is used to trackand analyze. Fish coordinates are extracted from the acoustic images, and fish aretracked by using an interacting multiple model (IMM) algorithm and joint probabilisticdata association (JPDA) algorithm. The swimming speed and direction of tuna can bepredicted and shown in real-time.Knowing the escaping velocity of tuna, we can try tocontrol the net promptly to prevent tuna escaping.The content of this paper is included with the sonar as main equipment, thecomputer algorithms, underwater detection and target tracking technology as mainmethods. Discuss the sonar imaging methods, the description of data receive, displaysystem, data format and its processing method. First, study the sonar data receivingapparatus connected to the data record system which is installed on the fish boats at sea; Secondly, study the sonar image processing methods according to the sonar dataprocessing method; Again, use3D processing technology to achieve the3D demo ofsonar underwater image data; Finally, use target tracking technology to explore the stateof motion of tuna fish in purse seines.The main research results in this paper are listed as follows:1. According to the imaging characteristics of the sonar, re-draw and preprocess thesonar image. The use of threshold filtering algorithm can significantly distinguishbackground noise and the target echo. This method can process multiple continuoussonar images fastly and automaticly, to facilitate subsequent target tracking research.Using Photoshop Illustrator ai, we can convert the single frame of sonar images intovector image data format, and import it to the3DMax software to process.2. When using the IMM-JPDA algorithm and Kalman filter tracking, we simulatethe tuna fish motion state with CV, CA and Brown model as its state transition matrix.Track100frames image of a data, the Results shows that tuna fish speed is within in0-10m/s. The maximum speed is35m/s.The minimum distance away from the net iswithin a distance of100m to300m range.From the speed and the distance, we canconclude that tuna fish movement state is changable in the period of seine fishing.3. In3DMax software we process the sonar image which has been converted intovector format, and export3D display video files.The sonar is also used to monitor theunderwater seine net and fish in3D Demo.Since the tuna fish is mainly floatinggroup,and the sonar horizontal beam emission angle changes from-4°to-45°, Thefish group was not obvious in the3D display.Purse seine net is displayed as the hugesound wall in the water.