Deep Transfer Learning for Automated Plant Disease Identification Using Mobile-Based Agricultural Imaging

Abstract

Plant diseases significantly reduce agricultural productivity and threaten global food security. Early disease detection enables timely intervention and minimizes economic losses. This paper proposes a deep transfer learning framework for automated plant disease identification using images captured through smartphone cameras. Pre-trained convolutional neural network architectures, including EfficientNet and MobileNetV3, are fine-tuned using publicly available plant disease datasets. Image enhancement and data augmentation techniques improve model robustness under varying environmental conditions. The proposed framework achieves high classification accuracy while maintaining low computational complexity suitable for mobile deployment. Comparative analysis demonstrates superior performance over traditional machine learning approaches. The developed system provides farmers with an affordable and intelligent tool for precision agriculture and sustainable crop management.