Onion Purple Blotch Disease Severity Grading: Leveraging a CNN-VGG16 Hybrid Model for Multi-Level Assessment

Abstract

Accurate and timely evaluation of Onion Purple Blotch Disease (OPBD) severity is highly important for efficient management of crop outputs and reduction of economic losses in onion cultivation. The old techniques, based on visual exploration, are time-consuming and the accuracy of inspection depends on the personnel carrying out the inspection. The current study begins a new way of using a Convolutional Neural Network (CNN) with a VGG16 architecture to automatically grade the multi-level severity of this disease. Through the model getting trained on a multi-labeled dataset of annotated onion leaf images, this study took a considerable step forward in the development of automated disease identification and classification of plant diseases. The best of the CNN-VGG16 hybrid model is shown in the overall accuracy of 93.5% which exceeds several modern models. This paper will outline the employed in data collection, model customization, and iterative training and validation. Besides revealing the model’s high precision in disease severity grading, these findings demonstrate the possibility of combining deep learning technologies in precision agriculture too. The research resulted in a new artificial intelligence (AI)based agricultural solution that fills the existing gap and can be successfully scaled up, rapidly identify the disease, and transform the agricultural technology landscape. Other directions involve the model deployment to crops and diseases and the integration of it with tracking aircraft to make the analysis better.

Topics & Concepts