Using image processing technology for water quality monitoring system

As fish has been existing and adapting to the water ecological environment that it will sense physically when water quality changes. Thus, the fish responding behavior has been taken one of the methods in monitoring water quality in recent years. This study has successfully in building a water quality monitoring system by utilizing the image processing and fuzzy inference in auto-recognizing the gesture of fish. It was our first time in setting up the image background model by using W4 method, and then adopted deduction of background in recognizing the fish profile. After finding the center-of-gravity position of fish profile, we can obtain the real time characteristic information of fish in position, speed and moving track. Finally put these information the input of fuzzy inference system, via appropriate rules bank in analyzing, the output value can be obtained. In this study, Zebra fish and Common Goldfish were adopted to be the study objects by using different into water and out of water device as well as different concentration of agent in observing the fish in response. From the result of experiment, the inferential method as proposed by this study in recognizing two kinds of fish has come to a satisfactory effect.