Plant Disease Identification using Machine Learning for Precision Agriculture

Abstract

The increase in the world population consequently prepares the need for food on a fundamental level. As a result, agriculture is important throughout the world. To meet human demand and provide money for farmers, a variety of crops, vegetables, fruits, fish, and animals are raised all year round. Nevertheless, occasionally individuals cultivating crops and grains suffer partial or even entire harm due to a lack of suitable cultivating knowledge, expertise, and awareness of disease prediction. Detecting disease at an early stage and providing suitable remedies to protect plants is a challenging task for farmers. Leaves are largely affected by several diseases, the majority of which are fungi, such as apple scab, Marconian coronary, black rot canker, powdery mildew, apple mosaic, and other viral infections. The experimental outcome has shown promising results of the proposed hybrid models over the recent state-of-art methods with a maximum precision of 99.3%, recall of 99.1%, accuracy of 99.8%, and F-score of 99.1%. In addition, the disease severity of the diseased leaf image class can also be identified by designing a novel optimized model with less computation time. newlineThe study includes an IoT-based monitoring system for real-time tracking of environmental factors like light, temperature, and soil moisture. Image preprocessing is done using contrast enhancement and histogram equalization, while dimensionality reduction is achieved through singular value decomposition (SVD). The system demonstrates improved disease detection and recommendation approaches, evaluated on monitoring datasets under various conditions. newlineThis research introduces a deep learning model for detecting plant diseases from images. The proposed model, implemented using Python, identifies healthy and diseased leaves and offers fertilizer recommendations based on soil conditions using an NPK sensor. newline newline