Daftar Isi:
  • This study aims to group tins based on RGB colors using K-Means Clustering. The effectiveness of this grouping is measured by accuracy level. The data used in this study consist of 250 sample of tin with several external factor on the tin images, which are four different lighting angels, two different velocity of Conveyor Belt, and two different lights. This study aims to find which external factor that gives highest accuracy and all grouping accuracy that the method gave. This study uses Randomized Group Design to find significant factors for the grouping. Therefore, for the red color the significant factors are velocity of conveyor belt 2, lights 1, and 450 of lighting angles. For the blue color are velocity of conveyor belt 2, lights 1, and 450 of lighting angles. velocity of conveyor belt 2, lights 1, and 450 of lighting angles This study use K-Means Clustering in order to group tin images based on RGB colors. The data used are the output data from randomized group design and 16 original tin’s RGB colors resulted from external factors combination. The K-Means Clustering method is using three kind of distance measuring which are Euclidian distance, Manhattan distance and Minkowski distance. The highest accuracy that the method gives is 49,6% from using Conveyor Belt 1, lights 1, and 450 of lighting angles, and by using Euclidean distance in the K-Means Clustering calculation.