Optimalisasi Identifikasi Buah Apel dengan Kombinasi Median Filter Dan Metode K-Means Clustering
A Afriadi, A Angga, Imelda Rosa, Windra Yosfand, Rini Sovia
Abstract
This study explores the application of the K-Means Clustering method to identify two types of apples, Fuji and Green apples, through image-based grouping by color characteristics. In this research, 20 image samples, consisting of 10 Fuji apple images and 10 Green apple images, were used to evaluate model accuracy. The identification process began with color space conversion from RGB to LBA, which proved effective in enhancing clustering accuracy, as the LBA color space better represents the visual characteristics that distinguish the two apple types. Additionally, median filtering was applied in the image pre-processing stage to reduce noise, significantly improving segmentation quality. Results showed that the K-Means Clustering method successfully identified all images accurately, achieving a 100% accuracy rate. However, further pre-processing techniques, such as lighting normalization, are suggested to improve model stability under varying image conditions. These findings indicate that combining K-Means Clustering with median filtering can be an effective and accurate solution for image-based visual identification.