A Model for Analysis of IoT based Aquarium Water Quality Data using CNN Model

Water is essential for the survival and wellness of humans and other ecosystems. The mix of physical and chemical characteristics in a water sample is referred to as water quality. Mainly, water quality is critical in achieving a long-term sustainable aquaculture system. Conducting water quality parameter evaluations is crucial for executing an assessment operation and developing a more effective water resources management and planning strategy in aquaculture. The cumulative impact of water quality parameters may devastate the whole system if not managed properly. Predominantly, early diagnosis of fish illnesses and identifying the underlying causes are critical for farmers to take precautions to prevent an epidemic. Typically, fish illnesses are caused by viruses and bacteria. The presence of these pathogens may alter the pH, D.O., BOD, COD, E.C., PO43-, NO3-N, and NH3-N levels in the water. Previously, water quality was measured using handheld devices and mathematical functions. But these models fail in getting practical water quality analysis. The advances of IoT and deep learning technology adopted to aquaculture give insight analysis on the water for aquaculture. Most of the researchers working with water quality data based on one parameter and analysis with limited samples. This paper used the convolution neural networks (CNN)to classify water quality data to classify the water quality effectively for fish growth and survival with all the water quality parameters and with a good number of samples. The proposed model is compared with the existing models. Performance results show the impact of the proposed model.