Key points are not available for this paper at this time.
Pollutants introduced into water sources by industries and individuals have made it difficult to obtain enough clean drinking water in the modern world. This calls for a mechanism that periodically assesses the water's quality. There are portable instruments available to measure the factors impacting water quality. But the user needs solutions that are inexpensive, automated, reliable, and responsive. For monitoring of large water bodies, a low-cost, scalable, edge processing system in real-time is created to satisfy the current needs. It is designed using a Raspberry Pi, a display unit, and standard sensors for variables including temperature, ORP, pH, conductivity, and DO. Edge intelligence along with tiny ML is used to deploy the virtual sensor for estimation of the degree of pollution in water which otherwise requires a costly and tedious process. With an edge reaction time of 2.70ns and a notable Nash-Sutcliffe efficiency coefficient of 0.99 for BOD and COD estimation in water, this model is created utilizing the randomized tree classification. With the use of a power-efficient algorithm, the 8.4w system energy consumption of the created prototype allows it to operate on AC and solar power and can be effective in difficult climatic conditions.
Aleefia .A. Khurshid (Thu,) studied this question.