Automatic Plant Disease Detection Using Deep Learning
Abstract: Plants get affected by different types of diseases. Each year a significant amount of food loss occurs due to various diseases globally. To ensure food security worldwide, the minimization of food loss due to plant diseases is essential. Diagnosing the diseases at the right time is crucial for effectively treating plants. Usually, the farmers or plant scientists diagnose the diseases. They perform the diagnosis by visually inspecting the leaves or different parts of the plants. This task is time-consuming and requires experience. Sometimes this type of manual detection process becomes more difficult due to the number of plants and the similarity of the symptoms. To alleviate this process, an automated plant disease detection model is developed using the Convolutional Neural Network (CNN) architecture of deep learning. The proposed model's computational expense and run-time are reasonably low. The model is validated using 2-D color images of potato leaves, including healthy and infected, collected from the PlantVillage dataset of the Kaggle public website. Test accuracy, the required number of parameters, and model run-time are considered to evaluate the model. The result is comparable to the state-of-the-art models to detect plant diseases with fewer resource requirements.
Loading