Abstract
The world’s second-largest producer and exporter of rice is India. Early disease detection is crucial to ensuring healthy rice production. In order to address the issue of diagnosing illnesses in rice plants, a number of strategies have been put forth; nevertheless, it has been discovered from the literature that these models do not perform with the anticipated accuracy. In this study, the best fit model among four CNN deep learning algorithms for categorizing rice leaf illnesses was attempted. Patterns were divided into four groups using 1600 images: healthy, brown spot, hippa, and life burst. Based on the findings, the learning rate, accuracy, and disease recognition accuracy of the performance comparison were examined. ResNet50, VGG19, InceptionV3, and ResNet152, CNN deep learning models, obtained disease recognition accuracies of 75.76%, 87.64%, 96.46%, and 98.36%, respectively. The classification efficiency result demonstrates that ResNet152V2 is the best CNN model for classifying diseases affecting rice.
Keywords: Deep learning, Image Classification, Leaf disease Detection, Rice Plant Leaf Disease identification.