Using low-quality video sequences for fish detection and tracking

This study demonstrates computational methods for the automatic detection and tracking of fish from video sequences. The research in this subject is very important especially in fish farming companies and for nature protection around the world. The process of automated control and counting of individual species of fish has a supportive contribution both in the nature conservation and in the food industry. The difficulties and the specific features of the problem have been studied in order to find a solution for the components of the automated fish tracking system. Typical methods in computer vision applications, such as background subtraction, Kalman filtering and Viola-Jones method were implemented for the motion detection, tracking, and estimation of fish parameters. Both the choice of the appropriate methods and results of the experiments strongly depend on the quality and the type of video sequences. The results we received from the practical experiments demonstrated that not all methods can produce good results for a real data set, whereas on a set of synthetic data they operate satisfactorily. These results suggest that the proposed approaches are adequate in estimating the annual amount of fish migrating through the fish pass.