Klasifikasi Jenis Kendaraan (Helikopter, Mobil, Motor) Menggunakan Metode K-Means Clustering pada Pengolahan Citra

Nurjannah, Farah
Ramadhanu, Agung

Jurnal Sains dan Teknologi
Vol. 5 No. 3 (2025): September-Desember

Digital image-based vehicle type classification still faces obstacles because the identification process is generally done manually, so it takes a long time and has the potential to result in object recognition errors. This condition indicates the need for an image processing-based automation system that is able to recognize vehicle types accurately and efficiently. This study aims to develop a vehicle image classification system (helicopters, cars, and motorcycles) using the K-Means Clustering method to improve identification accuracy based on visual characteristics. This study was conducted with a quantitative approach through four main stages, namely image preprocessing (RGB to LAB conversion and size normalization), segmentation using the K-Means Clustering algorithm, extraction of shape features (metric, eccentricity) and texture (contrast, correlation, energy, homogeneity) based on Gray Level Co-occurrence Matrix (GLCM), and evaluation of accuracy using a confusion matrix. The research dataset consists of 30 vehicle images divided equally for each class. The results show that the combination of the K-Means Clustering method and GLCM feature extraction is able to classify three types of vehicles with an accuracy level reaching 100%. These findings prove that the K-Means method is effective for vehicle image recognition automation, and can be used as a basis for developing artificial intelligence-based visual identification systems in the future.