Pengembangan Model Pembelajaran Mesin Prediksi Kematangan Buah Pisang Berdasarkan Citra Digital

Abstract

This study aims to develop machine learning to detect banana ripeness based on the skin color. In addition to the color of the skin, also carried out a comparison between the color of the banana with black spots on the banana. The color value used is RGB Color values are taken by using k – means to take the most dominant color value on the banana. After the data is obtained it will be learned using k - nearest neighbor. So as to produce a machine learning to get accuracy for this detection compared with 80% of the trained data and 20% of the tested data. So that the accuracy of the data that is only based on color is compared with the result value is 95.74% with the data using color data and the percentage of the color dominance value is 51.06%.