A Mobile Application for the Early Detection of Dolichos Yellow Mosaic Virus in Lablab-Bean Using Pre-Trained CNN Models and Deep Learning

Lablab-bean is a versatile legume crop that is widely cultivated across India. Lablab-bean is subject to various plant diseases on a global scale. Dolichos yellow mosaic virus (DoYMV) disease stands out as a significant limitation in Lablab-bean production across India. A rapid early detection system is essential for the timely management of this disease. This research examines the application of deep learning methodologies, especially utilizing pre-trained Convolutional Neural Networks (CNNs) for precise detection of virus symptoms on bean plants during the early stages of the disease. This study employed multiple CNN models, such as EfficientNet-B0, EfficientNet-B1, MobileNet, MobileNetV2, and MobileNetV3, to differentiate between healthy beans and those infected with the DoYMV. The results indicate that the MobileNetV3 models attain a remarkable accuracy of 99.97% in the training and $100 \%$ in the validation. The elevated accuracy rates highlight the promise of deep learning methodologies in advancing disease detection in crops, hence contributing substantially to the improvement of agricultural practices and the assurance of food security. This research emphasizes the significance of incorporating cutting-edge technologies for the early detection of crop health, aiming to reduce the effects of plant diseases at their initial stages of occurrence itself.