Research Of Fish School Acoustic Scattering Model And Its Simulation

The underwater acoustic technology is the most efficient approach used in fishery production and survey. Generally there are two main kinds of methods in the acoustic fishery application, which are measurement and model analysis. The acoustic scattering model of fish school is the base of acoustic fishery application. So, to have a deeper knowledge about fish school scattering problem and provide a theoretical support for the fish finder development, some efforts are made as follows:The single fish model is the base of fish school acoustic scattering model. Firstly two single fish acoustic models are given to describe the scattering mechanism of fish. We treat fish as the combination of body and swimbladder. In low frequency band, a sphere model is used for swimbladder, which is the main scattering contributor in this situation. While in high frequency band, both the body and swimbladder are modeled by Kirchhoff ray-approximation model, which takes into consideration the actual shape of fish. Based on the above discussion, an acoustic scattering model of fish school in low frequency is proposed with inclusion of the multiple scattering and coupling effect between each fish in fish school.The echo simulation of fish school is necessary to be developed because of the need to improve the accuracy of fish abundance estimation and fish classification algorithm. After the detailed research of fish model as well as the fish school scattering model, an exact simulation of fish school echo is done to generate the echoes in different fish school conditions.The statistical model is used as a tool to investigate the fish school echo characteristic. Some popular statistical models and approaches are applied to analyze the simulated fish school echo, and it is concluded that Gamma and K distribution fit the echo peak and envelope respectively.The fish target strength (TS) estimation is vital to the fish abundance survey. A single beam integral equation is used to model the acoustic fishery measurement. Two indirect estimation methods for this problem are particularly shown One is deconvolution and the other is Expectation-Maximization-Smoothing (EMS). It is found by simulation that EMS is a more promising estimation method over deconvolution in fishery estimation problem.