Automated Disease Detection of Rice Crops Through Video Recognition

Rice is a standard in over 50% of the world's population, yet its cultivation faces severe losses of 20-40% due to various diseases. Timely and accurate disease detection is crucial for effective management and higher yields. Traditional visual inspection is impractical due to scale and cost, necessitating advanced approaches. Presents efficient solutions based on deep learning. Use a deeper folding network and faster R-CNN combination to accurately identify and classify rice blade diseases. The video based detectors are the VGG19 based portion is learned by an adjusted transfer learning machine-learning approach comprising of grouped components that allude to a custom- made prosthesis backbone. Finally, the model is tested on rice leaf images and videos, surprising the results of VGG16 and YOLOv3 in terms of disease-classifiction accuracy and video-based detection, respectively. Finally, Mask R-CNN is investigated to identify more than one disease, achieving a maximum accuracy of 96% for blast, Brown Spot, and Sheath Blight. Millions of farmers can benefit from such technology and in light of that a survey of CNN techniques used for rice disease detection was thoroughly reviewed. These methods promise advances in disease management, improved yields of rice, and overall food security worldwide.