Artificial Neural Networks or Disease Diagnosis and Categorization in Citrus Plants

Plant diseases are the leading cause of lost production and productivity in agriculture, which in turn leads to significant economic expenses and food insecurity. Citrus is a plant genus that may produce fruit that has a high market value and is thus cultivated and harvested in large quantities all over the world. The cultivation of citrus, however, has been significantly hampered by the presence of a number diseases, which has resulted in significant decreases in both productivity and quality. The detection and classification of plant diseases has been greatly aided by the development of machine learning and computer vision in recent years. This improvement came about in recent years. Although machine learning models make use of a wide variety of methods for identification and prediction, it would appear that study into the deep learning (DL) area offers the most potential for boosting accuracy as a result of the confluence of these two fields. This research makes use of a two- stage deep Convolutional Neural Network (CNN) model in order to analyse leaf images for the purpose of identifying plant diseases and classifying citrus diseases. Using a region proposal network, the proposed model will first identify potential target diseased areas, and then it will use a classifier to assign the most likely target area to the right disease class. These two fundamental steps are what make up the model. The model that was proposed has a precision that is averaged at 96% and an accuracy of detection that is 95%.