A new approach to diagnose three grape plant diseases (Black Rot, Black Measles, and Leaf Bligh) based on color image processing and machine learning

Najafabadiha, Mohsen; Mohammadzamani, Davood; Gholamiparashkoohi, Mohammad

doi:10.22034/jrmam.2025.14711.698

A new approach to diagnose three grape plant diseases (Black Rot, Black Measles, and Leaf Bligh) based on color image processing and machine learning

Document Type : Original Article

Authors

Mohsen Najafabadiha ¹

Davood Mohammadzamani ²

Mohammad Gholamiparashkoohi ³

¹ Ph.D. student, Biosystems engineering, Islamic Azad University, Takestan branch, Takestan, Iran.

² Associate Professor, Biosystems Engineering, Islamic Azad University, Takestan branch, Takestan, Iran.

³ Department of Mechanical Engineering, Shahr-e-Qods Barench,Islamic Azad University, Tehran, Iran.

10.22034/jrmam.2025.14711.698

Abstract

Disease management in grapevine has become one of the basic issues that farmers have to face. Diseases can greatly affect the performance and life of this plant. With increasing environmental pressures and climate changes, the need for new and smart methods in managing diseases and pests has become inevitable. In this research, in order to diagnose and classify grape plant leaf diseases with the names Black Rot, Black Measles and Leaf Blight, after removing the background from the images of the leaves and extracting the features of texture, color and shape from the images, A combination of support vector machine classifier and butterfly optimization algorithm was used to select the most important features in the diagnosis of grape plant leaf disease. The results of the precision for black rot, Black Measles, and Leaf Blight diseases and healthy leaves were 100, 100, 100 and 95% respectively, and the classification accuracy for the diagnosis of the diseases and healthy group was 98.75%. Also, 15 characteristics of texture, color and shape were introduced to the researchers of plant pathology and data science with the help of butterfly optimization feature selection algorithm. The classification results showed that the use of image processing and machine learning has a high ability to diagnose and classify plant diseases.

Keywords

Feature Selection

Feature Extraction

Grape Diseases

Diagnosis and Classification

Artificial Intelligence

Subjects