Identification of Rice Leaf Diseases Using CNN and Transfer Learning Models

Rice is a crucial crop for global food security, but its production is often threatened by diseases that can significantly reduce crop yields. Detecting and diagnosing these diseases early on is crucial for managing rice crops effectively. The detection and classification of rice leaf diseases can now be automated due to the progress made in deep learning and computer vision technologies. The present study introduces an novel methodology for the identification and categorization of rice leaf ailments utilizing convolutional neural networks (CNNs), specifically the ResNet50 and VGG16 models. The proposed methodology employs deep learning methodologies to autonomously acquire features from images of rice leaves, thereby enabling the classification of various disease types. The proposed methodology is assessed through experimental trials on a dataset of rice leaf images that encompasses various disease categories. The findings indicate that the suggested methodology attains elevated precision in identifying and categorizing diverse varieties of rice leaf ailments. The ResNet50 framework surpasses the VGG16 framework in accuracy and sensitivity. The method under consideration exhibits the potential to serve as a proficient instrument for disease management in rice cultivation. It furnishes prompt and precise information regarding the disease status of rice crops, thereby empowering farmers to undertake essential measures to curb the spread of diseases and avert yield losses. Furthermore, the suggested methodology has the potential to be applied to various other crops, thereby facilitating the creation of disease detection and classification techniques that are more widely applicable and expandable.