Leaf Disease Detection in Chili Plants Using GLCM and HSV Combination with SVM Classification

Abstract

Chili pepper is one of the high-value horticultural commodities in Indonesia. However, this plant is highly susceptible to various leaf diseases such as yellow virus, leaf spots, leaf curl, nutrient deficiency, and whitefly infestation. Manual disease detection is often inaccurate and time-consuming, necessitating an automated solution that is more efficient and effective. This study aims to detect chili leaf diseases using texture and color features extracted from leaf images. This approach enables farmers to easily identify the type of disease affecting chili plants, allowing for faster and more precise control measures. The research utilizes 1,150 chili leaf images divided into five disease categories—yellow virus, leaf spot, leaf curl, nutrient deficiency, and whitefly—each consisting of 230 images (184 training and 46 testing data). Feature extraction is performed on color features using the Hue, Saturation, Value (HSV) color space and on texture features using the Gray-Level Co-occurrence Matrix (GLCM) method. For classification, the Support Vector Machine (SVM) with Radial Basis Function (RBF) kernel is employed. Parameter testing of C (1, 5, and 10) and Gamma (0.1, 0.01, 0.001, 0.0001, and 0.00001) shows that the best performance is achieved at angles 0° and 135°, with C=10 and γ=0.1, yielding a classification accuracy of 91.30%. These results indicate that the combination of GLCM and HSV features, along with optimal RBF kernel parameter tuning, effectively enhances classification accuracy.