ARCHIVES

Crop yield and fruit quality are both impacted by plant diseases, which are a serious threat to agricultural productivity. The commercially significant fruit crops papaya (Carica papaya L.) and mango (Mangifera indica) are particularly vulnerable to bacterial, viral, and fungal infections. Manual visual inspection, which is frequently laborious, subjective, and reliant on expert knowledge, is the foundation of traditional disease diagnosis. This research proposes a deep learning-based framework for automated disease classification utilizing fruit photos in order to overcome these constraints. EfficientNetB0's outstanding classification performance and processing efficiency make it the foundation of the suggested solution. Anthracnose, Black Spot, Ring Spot, Phytophthora, Powdery Mildew, and Healthy Papaya are the six classes in the papaya dataset; Alternaria, Anthracnose, Black Mould Rot, Stem End Rot, and Healthy Mango are the five classes in the mango dataset. To improve model resilience and generalization, image preprocessing methods like scaling, normalization, and data augmentation are used before training. The algorithm automatically extracts disease-specific features and conducts multi-class classification of fruit samples that are both healthy and infected using transfer learning. By lowering crop losses and promoting precision agriculture, the suggested framework seeks to offer an accurate, scalable, and economical approach for early disease identification. Additionally, the approach might help farmers and agricultural specialists make prompt judgements on disease control, which would increase crop sustainability and output.