A method to estimate the abundance of fish based on dual-frequency identification sonar (DIDSON) imaging

In this paper, a new method based on the dual-frequency identification sonar imaging is presented to estimate the abundance of fish. The proposed method can detect, track and count fish individually. Fish targets can be acquired with a depth-first search algorithm once its edge is detected based on the Sobel Operator from a complicated image. Multiple fish targets can be tracked using the nearest neighbor algorithm combined with the Extended Kalman Filtering. In order to validate the effectiveness of the proposed method, an experiment was conducted in Qingcaosha reservoir in Shanghai. The manual detection method based on sonar images was used as a reference method. The results showed that the fish were identified with less than 5% error compared with the manual detection results. The statistical error of fish numbers between the automated method and the reference method was approximately 10%.