Target Tracking And Early Warning Analysis Of Bio-style Water Quality Monitoring

Bio-style water quality monitoring is a very important research in field of water quality monitoring. Bio-style water quality monitoring technology compared to traditional physical and chemical water quality monitoring technology has the advantages of high sensitivity, low cost, intuitive and the strong representation of comprehensive analysis. Nationally and internationally, it has been widely studied and applied in recent decades.This paper applied motion detection and tracking technology based on video to the fish water quality monitoring system, then assessed the quality of water by measuring, analyzing and comparing characteristic changes of the behavior of the zebrafish in different environmental conditions. This paper researched video target detection and tracking technology which is used in fish water quality monitoring system, using the fish’s trajectory by video tracking to calculate the basic behavior parameter data of live fish, these behavior data changes can be used to understand the behavior of the changes of the live fish, to achieve the purpose of qualitative evaluation of water quality.The main work in this paper includes four aspects as follows:1、Video detection and tracking technology, water quality monitoring technology, and some image processing are researched.2、The video detecting and tracking methods of the fish in the live-fish water quality monitoring system are researched. In this paper, we get the initial background using a statistical background modeling method, updates the background during target detection and tracking in real-time. On this basis, background subtraction method and adaptive image binarization are used to achieve the extraction of fish moving target, and the connected component analysis is used to extract fish size, mass and other characteristics of value, and finally we applied the Kalman predictor for the motion tracking based on characteristics to track the fish’s trajectory.3、The early warning methods of the live-fish water quality monitoring system are researched.In this paper, based on the behavior of fish, we set the average swimming speed, the speed index, the average swimming height, the average swimming distance and the number of fishes as the basic parameters. These parameters can be calculated using the fish’s trajectory extracted by video tracking, and then the change of live fish’s movement state is detected by the threshold method and Hinkley detector, achieving the goal of water quality’s early warning.4、The live-fish water monitoring software is developed by the C++programming language using the Matrox’s Image processing library of MIL8.0. The developing tools is provided by the Visual Studio2008. Then, the experimental environment is established, so it can be used to process and analyze every frame of the video sequences. It is realized that the function of the live water monitoring basically.