Aquaculture Monitoring System: A Prescriptive Model

Aquaculture is one of the fastest-growing industries in the world. Billions of people depend on it for food and livelihood. Yet fishermen are reluctant to adopt aquaculture as it is expensive and difficult to manage and monitor. Minute changes to the environment can significantly affect the fish, causing suboptimal growth, and in extreme cases, leading to the death of fish. Economic loss due to such changes is a major deterrent for people adopting aquaculture. Most of these losses can be avoided as these minute changes can be prevented by taking simple countermeasures. However, the fishermen are unable to take these countermeasures as they are unable to assess the situation due to a lack of easy and real-time access to the current condition of the ecosystem. The length of a fish is an excellent indicator of its health. Traditional methods of measuring fish length involve removing the fish from the water and manually measuring it, which is inefficient, inaccurate, and stressful for the fish. Advancements in technology have not only made it possible to monitor the current condition of the ecosystem, but also to predict future conditions. This paper analyzes the effect of water quality parameters on fish growth. An LSTM-based model is used to predict these parameters. A prescriptive model that advises fishermen is built on the predictive model. A monitoring system that provides easy and real-time access to the water condition and the predictions is proposed. This paper also explores the application of stereo vision for non-contact estimation of fish length, enabling the fishermen to better assess the situation and make better decisions.