Vgg-Resnet Hybrid (VRH) Model for Automated Classification of Cotton Leaf Diseases

Although the development of advanced fluorescent techniques has been occurred in the last few decades, the automation in plant disease detection has become a key research area for sustainable agriculture in recent years. A new model architecture known as the VGG ResNet Hybrid (VRH) is developed based on the merits of the VGG16 and ResNet50 models for the classification of cotton leaf diseases. Given that the mechanized care model relies on well-developed CNNs, fine- tuned on a developed cotton leaf disease digital dataset, the strategy is designed to improve classification performance as compared to a traditional set of Machine learning supervised classification algorithms. To understand the performance of the proposed VRH model, the model is compared with other pre- trained models like VGG, ResNet and Inception to analyze the success of these models for identifying cotton leaf diseases. The independent models themselves are compared against the generated hybrid model, and as seen in figures above, accuracy and efficiency are greatly enhanced. Deployable in crops, the study demonstrates the ability of fine-tuned pre-trained models in early disease identification and an efficient solution to assist farmers in timely interventions hence reducing crop loss, rational use of pesticides and achieving sustainable agriculture.