DEEP LEARNING ALGORITHMS FOR RICE LEAF DISEASE DETECTION

Abstract

  Rice cultivation is integral to global agriculture, and the identification of diseases affecting rice plants is crucial for ensuring optimal crop yield and quality. This research presents a system that utilizes computer vision and techniques of deep learning to automatically detect diseases in rice plants. The study evaluates various machine learning and deep learning algorithms for their accuracy, recall, and precision in identifying common diseases of rice leaf such as “brown spot, leaf blast, bacterial leaf blight” and others and emphasizing the superior performance of deep learning models. The EfficientNetB0 model, fine-tuned with data augmentation, demonstrates remarkable accuracy at 98.2%, showcasing its potential for practical applications in precision agriculture. Future work involves expanding the dataset, exploring alternative model architectures, and implementing real-time disease detection in the field using edge computing or mobile applications. This research contributes to advancing agricultural technology and intelligent crop disease management, marking a significant step towards sustainable and efficient farming practices.