Agroecology: Intelligent Implementation of Digital Farming Using machine Learning & Deep Learning

Agroecology is one of the revered and illustrious professions in India. Agribusiness occupies more than 60% of the country’s land, which feeds 1.428 billion people. The cultivation of plants and animals is known as agriculture. Farmers face a lot of hardship while using conventional farming techniques and also face heavy losses every year. Digital farming helps increase productivity by precisely determining the steps that need to be practiced at the appropriate time. It can be done in a laboratory, but it consumes a lot of time. That’s why nowadays researchers are focusing on digital farming techniques because it takes less time and give better results. Till date, many studies have been conducted to improve agricultural planning. In this research, we present an intelligent system called Digital Farming, which intends to assist the farmers in making an informed decision about which crop to grow, the fertilizer needed to improve the productivity of soil, and the water supply needed according to the soil. Not only the soil type and crop type play an important role in crop yield, but pests and weeds present in the agricultural fields also lead to heavy losses to the farmers, so the proposed model also detects the pests and weeds present in the field and also recommends pesticides. The work uses Convolution Neural Network (CNN) for soil classification, crop recommendation and fertilizer recommendation, for pest detection and pesticides recommendation it uses K Nearest Neighbour (KNN) for weed detection it uses Support Vector Machine (SVM). CNN shows 91% accuracy SVM shows 71% accuracy and KNN shows 83% accuracy.