A Method of Detection and Classification of Plant Diseases Using Classifiers

Abstract

ABSTRACT. Agriculture is one of the main sources of income for the government of Saudi Arabia. It is included in one of the twelve pillars of the 2030 vision. The Public Investment Fund which is also known as PIF has launched its strategic plan for 2021-2025 to thrive the economy for diversity. This strategic plan consists of numerous key initiatives and one of them is food and agriculture. It focuses on enabling the growth of food and agriculture inside Saudi Arabia to reach the sustainability level of domestic food production, ensuring food strength and flexibility and having a diversity food supply sources from outside the country. Hence, having disease-free plants is a must for this initiative. In this paper, proposing an algorithm to detect diseases in plants is presented. It works based on combining the K-means and Convolutional Neural Networks (CNNs) approaches. The obtained results show that this method achieves more than 95% of accuracy. The simulation experiments are conducted using MATLAB as a simulation tool. Finally, an assessment between the proposed algorithm and some state-of-the-art works from the literature is presented. This evaluation shows that the implemented method outperforms other works in terms on accuracy, precision, and recall respectively.