Harnessing deep learning algorithms for early plant disease detection: A comparative study and evaluation between SSD (Mobilenet_v2 and Mobilenet_v3) and CNN model

Abstract

Recognizing the important need for efficient plant disease detection in agriculture, this research evaluates and compares the  performance of three distinct deep learning models: Mobilenet_V2, Mobilenet_V3, and a custombuilt CNN model. As traditional methods  fall short in addressing the evolving challenges of crop health management, the study aims to discover the most effective model for  accurate disease identification. Leveraging a dataset encompassing 20,639 images across 15 directories representing various plant  diseases, the models undergo rigorous training and evaluation. Results reveal the CNN_model as the superior performer with a remarkable test accuracy of 94.48%, outshining Mobilenet_V2 and Mobilenet_V3. The comparative analysis sheds light on the strengths  and weaknesses of each model, providing valuable insights for the agricultural community. This research not only advances the  understanding of deep learning applications in precision agriculture but also lays the foundation for future innovations in sustainable  crop management.