Predicting Potato Leaf Diseases with Convolutional Neural Networks

Abstract

Early disease diagnosis in potato leaves is complicated by the wide range of crop types, agricultural disease signs, and environmental factors involved. These problems make it difficult to detect diseases in potato leaves at an early stage. For the purpose of identifying diseases in potato leaves, a number of machine-learning techniques have been developed. The models used to detect crop species and agricultural illnesses are only tested on photographs of plant leaves from a specific geographical area, limiting the effectiveness of existing methodologies. The farmer can prevent severe financial losses by promptly detecting and controlling such outbreaks. The results of this study contribute to a unique approach that makes use of image processing to accurately detect illnesses in potato leaf populations. There are several machine learning methods for spotting symptoms of disease in potato leaves pictures; here, we employ the Convolutional Neural Network (CNN) model. The goal of this research is to develop a convolutional neural network (CNN)-based sequential model for disease prediction in potato leaves. This study's model accuracy was 92.58%. The presented model was tested on both typical and deformed potato leaves, with mixed results. Next, the algorithm is applied to the images, and the potato tree's leaves are classified as healthy or unhealthy.