AGRIGURU: A smart artificial intelligence solution for crop recommendation and plant disease detection
TL;DR
The proposed system resolves the challenges by utilising an integrated software platform that delivers timely and relevant information to farmers, thereby improving agricultural productivity and performance, and provides farmers with timely and actionable insights.
Abstract
Agriculture plays a crucial role in influencing the economy of any country, as it is the major source of food, raw materials and employment to the majority of the population. In India, agriculture contributes a major portion to the country’s gross domestic product (GDP) and hence it is crucial to enhance the agricultural yield using the latest tools and technologies. Agricultural yield depends on factors such as plant diseases, fertilisers used, improper crop choices and inaccurate yield prediction techniques. Information Technology (IT) is growing rapidly and is utilised in every walk of human life. Recent advances in artificial intelligence (AI) and machine learning (ML) have enabled data-driven solutions to address these challenges in agriculture. Hence, this work targets to design and develop a system that identifies the infected plants, detects the type of disease and performs crop recommendation. The crop recommendation system uses the support vector classifier (SVC) as the core algorithm. The convolutional neural networks (CNNs) that are trained and tested with the Plant Village dataset, have given higher performance in identifying the disease types of plants, compared to K-nearest neighbours (KNN), Logistic Regression and Decision Trees. This recommendation system determines the nitrogen, phosphorus and potassium (NPK) contents, the pH value and the moisture percentage levels of the given soil and recommends appropriate crops for farmers. Soils from different regions have been considered for experimentation. The decision tree algorithm has been utilised for crop yield prediction by processing climatic and historical agricultural data. Remarkable precision has been achieved with the operations. The proposed system resolves the challenges by utilising an integrated software platform that delivers timely and relevant information to farmers, thereby improving agricultural productivity and performance. The integrated software platform provides farmers with timely and actionable insights, supporting informed decision-making and improved agricultural productivity.