Computer Vision-Based Model for Classification of Citrus Fruits Diseases with Pertinent Image Preprocessing Method

In the agricultural industry, diseases that harm plants are mostly responsible for the reduced profit that results in monetary loss. Nowadays, approaches based on computer vision and image processing are often used for identifying and categorizing plant diseases. The major goal of this paper is to offer an automated system for correct disease prediction at an early stage so that appropriate action can be taken at the right time and loss can be minimized because manual prediction is time-consuming. As a result, we proposed a classifier based on deep convolutional neural networks (DCNN). It has four primary steps, i.e., 1. preprocess captured images, 2. segment infected images, 3. feature extraction, and 4. classification. Firstly, preprocessing is ended to lift the image’s eminence by histogram equalization. To reduce overfitting, a data augmentation approach is used to increase the dataset. As a feature extractor, the AlexNet model is then used. The numerous types of citrus illnesses are finally classified using the proposed model and is validated with baseline classifiers like Support Vector Machine and Random Forest. The proposed model gives 98.26% test accuracy, which is better than other classifiers and is very useful for farmers.