Biological Early Warning System for Prawn Aquiculture

Abstract Water quality monitoring through observing physiological or behavioral changes of living organisms has been arouse much attentions from academia and industry during past years. In this study, a biological early warning system was designed to record prawn behavior characteristics subjected to different water qualities. The system consists of a specially-designed water container, a digital video recorder and image processing software. Prawn samples were divided into three groups for calibration and one group for validation. The targeted water quality parameters includes water salinity, pH and dissolved oxygen concentration (DOC) that were altered by adding different chemicals in the experimental water container. The position of a prawn was marked by its dark feces in its body. The movement of the prawn was then tracked by its position-time sequence, which allows calculating the moving speed of the prawn. The calibration result shows that DOC changes make great impact on prawn behavior while water salinity and pH do not. When DOC is above 3.0 mg/l, prawns move at about 80 pixels/s; when DOC is between 2.6 and 1.74 mg/l, prawn moves at about 120 pixels/s; and when DOC is below 1.2 mg/l, prawns move at the speed of 1000-1200 pixels/s. The validation set of prawns with DOC changing from 1.82 to 0.54 mg/l confirms the calibration observations. It is concluded that the proposed approach is promising for water quality monitoring, especially DOC in aquiculture environment.