Mango leaf disease detection using deep feature extraction and machine learning methods: A comparative survey
Yavuz Ünal, Muammer Türkoğlu
Abstract
Plant diseases significantly affect the quality and quantity of agricultural production. Diseases seen in the leaves of plants adversely affect plant growth and yield. In the near future, accessing cheap and safe food will be one of the most important problems of countries. Therefore, early detection of plant diseases is very important in terms of economy and access to food. It is very difficult to visually detect and monitor the diseases in mango leaves. This study aims to detect diseases in mango leaves with the aid of image processing and deep learning. Deep features are extracted from mango leaf images (by using Darknet19, Xception, SqueezeNet, MobileNetv2, DenseNet201, GoogleNet, ResNet18, VGG16 and AlexNet architectures) and classified with Decision Tree, Linear Discriminant Analysis, Naive Bayes, Support Vector Machine, k-Nearest Neighbors, Ensemble Classifier. As the results of the evaluations, it is observed that the results found in the literature were improved. Details of experimental results are presented in the article.