PLANT LEAF DISEASE DETECTION

Plant diseases significantly reduce agricultural yield and quality, making early and accurate detection essential for sustainable farming. This paper proposes a smart plant leaf disease detection and severity analysis system using a hybrid deep learning framework. The approach integrates advanced image preprocessing techniques with convolutional neural networks (CNNs) and transfer learning models to enhance disease recognition under real-field conditions. Preprocessing steps such as noise reduction, color normalization, and adaptive segmentation improve the visibility of disease symptoms. Deep features extracted from pre-trained CNN architectures are used for multi-class disease classification and severity estimation. In addition, a preventive advisory module provides actionable recommendations for disease control and crop management. Experimental evaluation demonstrates high accuracy, robustness to environmental variations, and effective generalization across different crops. The proposed system supports intelligent decision-making and offers a scalable solution for modern precision agriculture.