IoT based prediction of water quality index for farm irrigation

Agriculture sector of Indian economy in which more than half of the population is involved, contributes to less than quarter of the GDP. With advancement in ICT, tools and techniques can be developed that can help in analyzing and automating various phases of farming for improving productivity. This work focuses on analysing the quality of irrigation water and developing a model for prediction of Irrigation Water Quality Index(IWQI) based on Salinity and Sodicity. Development of IWQI can save time and cost of lab tests for irrigation water. Five parameters of water Na⁺, Cl^-, EC, HCO^3- and SAR are measured using which the IWQI is calculated. These five parameters are further reduced to three parameters using correlation analysis and a classification model for prediction of water quality class is developed using various classification techniques. Best result is obtained by Random Forest Classifier followed by Gradient Boosting and Neural Network Classifier. The classification model can be used in IoT based farming systems for preventing salinity based damage to the crops.