Budi Sunarko, Muhammad Rozy, Apriansyah Wibowo, Yudha Andriano Rismawan
The research has developed a Content-Based Image Retrieval (CBIR) system to detect the presence of canker and black-spot diseases on citrus fruits, using the Hue, Saturation, Value (HSV) color feature extraction method. The dataset consists of 50 images of citrus fruits for the collection and 30 images of citrus fruits for the query set at a 1:1 ratio among classes. Standardization of the images was performed by resizing them to 800x800 pixels. HSV histograms were extracted from the images as the features for measuring the similarity between the query image and the collection image with the Chi-Square distance. The CBIR system based on HSV achieved an average retrieval accuracy of 88% and provided high values of precision, recall, and F1-score, indicating the ability of the system to effectively distinguish between the two diseases. Therefore, it may be concluded that the HSV-based CBIR system may be utilized as a useful, practical, and understandable tool to support the classification of citrus diseases. © 2025 IEEE.
Universitas Negeri Semarang, Department of Electrical Engineering, Semarang, Indonesia; Institut Teknologi Sepuluh Nopember, Department of Statistics, Surabaya, Indonesia
