Image Processing And Machine Learning Approach For Tomato Leaf Disease Detection

Abstract

Using image processing and machine learning, this research presents a technique for quick plant disease diagnosis in tomatoes. Gray-level co-occurrence matrices (GLCM) are used in the proposed approach to extract textural aspects of leaves and histogram color extraction. The collected characteristics are then utilized to train a support vector machine (SVM) to categorize different communicable and non-communicable plant diseases. The trained model is put on a Raspberry Pi, which captures video using a pi camera and employs the classifying model to diagnose plant disease in real time. On the plant disease SVM model, which diagnoses two common illnesses, bacterial spot and mold, the generated system has an accuracy of 97.29%. It was tested on a few example leaf images to see how accurate it was in real time.