CNN-VGG16 Hybrid Model for Onion Purple Blotch Disease Severity Multi-Level Grading

The purpose of this study is to introduce a novel strategy for an automatic disease grading system that has been devised for the onion purple blotch disease which prevails in many farmers' fields. In identifying patients' level of disease the overall accuracy was about 96.7% using the hybrid Convolutional Neural Network (CNN)-VGG16 model. Image Preprocessing technologies that are advanced, Normalization as well as augmentation provide high models' reliability and permit the distinction between various degrees of disease expressions, ranging from trivial to severe. The model's accuracy, granularity, as well as its reliability to identify disease severity, are evaluated using comprehensive evaluation metrics consisting of a detailed confusion matrix and performance parameter results, including TP, FP, FN, and so on. The results relate to changing detection technologies and a higher emphasis on enhancing data quality as well as adjustments to specific environments. This study provides grounds for further developments such as multi-sensor data fusion methods and online integration of field data. These undertakings aim at ensuring improved precision of intervention and speed, providing farmers with a reliable and efficient means of quick diagnoses on purple blotch diseases, thereby promoting the sustainability of crops and yield enhancement in agriculture.