Hybrid Deep Learning Framework for Cashew Leaf Disease Classification Using Model Concatenation

In this paper, we propose a novel deep learning-based method for the accurate classification of cashew leaf diseases, utilizing a hybrid MobileNet-VGG19 Con concatenated model trained on the CCMT cashew disease dataset. Detecting cashew leaf diseases is particularly challenging due to the wide variability in texture, color, and structural patterns. To overcome these complexities, our model integrates the lightweight efficiency of MobileNet with the deep feature extraction strength of VGG19, achieving a powerful balance between computational speed and representational depth. The proposed model outperforms its standalone counterparts, achieving impressive results: 98.20% training accuracy, 96.23% test accuracy, 96% precision, 95% recall, and an F1-score of 96%. Beyond accuracy, the model demonstrates strong robustness and generalizability, making it highly suitable for real-world applications in precision agriculture. Our findings highlight the potential of hybrid deep learning models to revolutionize plant disease detection, supporting sustainable, automated, and intelligent crop management practices.