Classification of Orchid Types Based on Leaf Images Using the Gray Level Co-occurrence Matrix (GLCM) and K-Nearest Neighbor (KNN) Methods

Abstrak

Orchids are popular ornamental plants with high economic value. Various types of orchids in Indonesia are difficult to distinguish when they are not in bloom, especially for laymen. This study aims to develop a classification system for orchid species based on leaf images using the Gray Level Co-occurrence Matrix (GLCM) and K-Nearest Neighbor (KNN) methods. GLCM texture feature extraction includes contrast, correlation, energy, and homogeneity, which are then classified using the KNN algorithm. The research dataset consists of 88 leaf images for training data and 22 images for test data from 11 orchid species, namely dendrobium, arachnis, cattleya, arundina, phalaenopsis, cymbidium bicolor, doritis polcherrima, appendiculata, flickingeria aggulata, phaius tankervilleae, and vanda douglas. The test results show that k=1 produces the highest accuracy of 77.2%, k=3 produces an accuracy of 72.7%, and k=5 produces an accuracy of 59.0% on the test data. The selection of the optimal k parameter was performed using Grid Search with the cross-validation method to ensure better model generalization. This study shows that the combination of the GLCM and KNN methods is effective for classifying orchid species based on leaf image characteristics with good accuracy.

Identifier

• DOI: https://doi.org/10.24167/sisforma.v12i2.14315
• OpenAlex: https://openalex.org/W4416455462