A Relative Analysis for Plant Disease Detection with AI-Driven Techniques: Optimizing Precision Agriculture

Vegetable crops play a vital role in ensuring global food security but are highly susceptible to a wide range of diseases, posing serious challenges to sustainable agriculture. Early and accurate detection of these diseases is essential for maintaining crop health and maximizing yields. In this research, four advanced deep learning models such as EfficientNetB0, InceptionV3, DenseNet121, and MobileNetV2 are evaluated focusing on their ability to identify plant diseases from images. EfficientNetB0 stood out as the top-performing model, achieving a validation accuracy of 96.4% and excelling across key metrics such as precision (94%), recall (96%), and F1-score (94.8%). Its success lies in its compound scaling method, which optimizes the model’s depth, width, and resolution, making it both accurate and efficient. MobileNetV2, though less accurate at 92.3%, is a solid alternative for environments with limited computational resources. InceptionV3 and DenseNet121 performed reasonably well, but their accuracy and training efficiency fell short compared to EfficientNetB0. Overall, EfficientNetB0 proved to be the most balanced and reliable model, offering strong performance and adaptability. Its ability to detect diseases accurately, while minimizing false positives, makes it well-suited for practical applications in agriculture. The efficiency of the model allows it to be deployed directly in the field using low-power devices, providing farmers with real-time disease detection capabilities. This work contributes to the growing field of technology enhanced agriculture by offering a reliable solution for early disease detection, ultimately improving crop protection and ensuring sustainable agricultural practices.