Classification of Seagrass Types Using GLCM Feature Extraction and K-Nearest Neighbor (KNN) Algorithm

Abstract

Seagrass is a type of flowering plant (Angiospermae) that grows fully submerged in shallow coastal waters and estuaries, playing a vital role in marine ecosystems. Currently, seagrass species identification is still performed manually by experts, which is time-consuming, costly, and labor-intensive. To support more efficient conservation and ecological monitoring, an automated, fast, and accurate method is needed. This study proposes the combination of the K-Nearest Neighbors (KNN) algorithm for classification and Gray Level Co-occurrence Matrix (GLCM) for texture feature extraction. The seagrass image data was obtained from the Roboflow website, and the value of k used in KNN was set to 3. Feature extraction using GLCM was conducted at angles of 0°, 45°, 90°, and 135°. The results showed the highest accuracy at k=3, with 77.42% accuracy on training data and 73.33% on testing data. Therefore, the combination of KNN and GLCM has proven capable of providing fairly accurate results in identifying seagrass species.