Ai-Powered Crop Monitoring for Early Detection of Paddy Leaf Diseases Using Attention-Based Densenets

Abstract

Early identification of diseases, pests, and nutrient shortages in paddy crops will help to determine the degree of agricultural output. Considered traditional monitoring methods rely on hand inspection, which is not only time- consuming but also maybe erroneous and labor-intensive. Including AI (AI) into precision farming increases disease detection efficiency and helps to reduce crop losses. This work offers an AI-driven approach to detect paddy leaf disease using Attention-Based DenseNets. This approach makes advantage of image recognition, sensor data, and advanced machine learning techniques. Preprocessing, feature extraction, and classification come in three basic forms involved in the methodical process. The first step of the process meant to improve the quality and eliminate noise using which Residual Support Vector Machines (Residual SVM) help us to perform feature extraction so guaranteeing the choice of strong features from diseased leaf samples. Attention-based dense networks enable classification. These kinds of networks enhance feature representation and last result in accurate disease detection using attention mechanisms. Split into healthy and sick groups, the dataset contains 5,000 pictures of paddy leaf. The model is evaluated as well as trained using this dataset. Higher than some of the more conventional methods, CNN (92.4%) and ResNet-50 (95.1%), the results of the experiments show that the proposed model achieves a classification accuracy of 98.6%. Precision is 90-8%; recall is 90-2%; the F1-score for disease detection is 90-8%. Moreover, early identification made possible by the system helps to reduce the 42% disease spread using which crop yield prediction is enhanced.