Abstract:
"The spread of diseases in crop production is a major challenge to global food security. Among these crops, black pepper is sensitive to a variety of diseases, which can reduce productivity and quality. Using deep learning techniques, this research addresses the challenging task of detecting disease in black pepper plants in real time and with accuracy.
The author used the MobileNet architecture to create a smart black pepper disease detection system since it is efficient and adaptable to the specific purpose. Key improvements were made to the model, beginning with the selection of a deep layer for feature extraction, followed by changing its output to be compatible with a recently added classification layer aimed to distinguish between three disease stages.
The model performs extremely well in disease prediction, with a 99.87% accuracy rate. It is particularly good at diagnosing common diseases in black pepper plants, such as leaf blight, and yellow mottle. This breakthrough significantly improves agricultural technology by providing a reliable method for diagnosing problems in black pepper plants. Future attempts to develop the model include increasing data collecting and augmentation efforts. Furthermore, efforts will be made to provide accessible, user-friendly interfaces to encourage widespread usage and acceptance in the agriculture industry."
