Evaluation of The Deep Learning Techniques to Identify Plant Diseases Using Leaf Images

Abstract

Successful farming is influenced by various techniques conducted by farmers in identifying the types of diseases that affect plant yields to avoid greater losses. Therefore, this study aims to evaluate the deep learning techniques in identifying plant diseases using leaf images. Furthermore, an artificial intelligence approach was used to identify types of plant diseases. During the deep learning training, about 11 deep learning architectural models and consisting of 38 classes in the dataset were used. The results showed that the highest minimum accuracy value obtained was 87.10%, with only one class having an accuracy value below 90%.