Smart Paddy Identification and Soil Suitability Prediction Using Deep Learning

Abstract

Agriculture is vital in various aspects including economic stability and food security, especially in areas where paddy cultivation is the “predominant farming activity. To achieve high levels of productivity and viability in farming, identification of paddy varieties and evaluation of the suitability of the soil are important. Traditionally, identification and evaluation methods are labor intensive and time consuming and their results are uncertain due to the possibility of errors. In this regard, the study proposes an innovative and intelligent system based on the use of deep learning and machine learning to deal with the identification and evaluation problems. The innovative system utilizes deep learning to deploy Convolution Neural Networks (CNN) to identify paddy crops and analyze the conditions and characteristics of the crops. In the same way, paddy crops and the suitability of the soil are evaluated and appropriate recommendations are made to the farmers. The innovative system integrates image analysis and soil parameter evaluation to offer a comprehensive decision support system to farmers. A web-based support system is designed to offer a user friendly support system to enable farmers to conduct real time assessment of the crops and soil. The innovative system eliminates the need for manual evaluations and optimizes the resources used and the time taken to identify potential farming problems. In the same way, the innovative system designed to support management of agriculture in a sustainable way decreases the need for fertilizers and supports informed decisions when farmers practice farming.