Pesticide Recommender System for Detecting the Paddy Crop Diseases through SVM

Rice cultivation is pivotal in India, facing various color-related issues during different growth stages. Identifying these conditions manually can be challenging, especially for less experienced farmers. Recent advancements in Deep Learning (DL) demonstrate the effectiveness of automated image recognition systems. These systems utilize Convolutional Neural Networks (CNN), Transfer Learning (TL), and Support Vector Machine (SVM) models for similar agricultural problems. Due to the unavailability of readily accessible rice disease image datasets, a custom dataset was created. Consequently, a deep learning model was devised using transfer learning. The proposed CNN architecture, based on MobileNet, was trained and tested using datasets from paddy fields and the internet.