Automated Brinjal Leaf Disease Detection Using Machine Learning: A Sri Lankan Study

Abstract

The spread of plant diseases has a significant negative impact on agricultural output quality and yield, resulting in significant economic losses. Plant disease automation in agricultural science is our country's main concern due to the economic crisis and the emphasis on organic farming practices. Furthermore, the accuracy and efficiency of recognizing sick leaves have increased due to the expanding usage of smart technology. This study uses machine learning and image processing to create a system that can classify diseases in brinjal leaves with greater accuracy. For comparison with healthy leaves, tobacco mosaic virus, collar rot, bacterial wilt, and cercospora leaf spot were chosen. For this investigation, pictures of 35 leaves from the Tirunelveli Purple Brinjal (Solanum melongina L), including healthy leaves, were taken on-site in Atchuveli, Northern Province of Sri Lanka. First, using the KNN approach, Histogram Oriented Gradients (HOG) and Principal Component Analysis (PCA) based features were used to classify the images, with accuracies of 0.95 and 0.97, respectively. Increased training data is required to achieve high accuracy with these techniques. The suggested system will enable farmers to recognize the appropriate disease. The development of the full system, including the hardware, camera, and classification scheme that will be applied in the field, is the study's ultimate objective. In the future, we plan to design a complete system with automatic detection.