Identifikasi mutu buah naga putih (Hylocereus undatus) menggunakan pengolahan citra (identification of white dragon fruit quality using image processing)

Abstract

Dragon fruit is one of agricultural commodity which has good agribusiness prospect. The general objective of this research was to study the possibility of identifying the quality of the white dragon fruit (Hylocereus undatus) using image processing approach. Analysis of quality parameters of image processing was conducted on the length, diameter, area and intensity of the RGB color of the object. The results of calibration showed that the quality parameters which have the highest accuracy on identifying fruit weight and diameter were identification using parameters area and diameter of the image. The two highest accuracy obtained were at 84% and 54%. While identification of quality grade A, B and C using image area produced accuracy at 82%, 50% and 78%, respectively. While identification of quality grade A, B and C using diameter of the image produced accuracy at 68%, 38% and 76%, respectively. While identification of quality grade A, B and C using area and diameter of the image produced accuracy at 64%, 46% and 90%, respectively. The number of pixels for weight and diameter of grade A was 33.402 and 184 pixels; for grade B 29.868 - 33.402 pixels and 173 - 184 pixels; to: grade C < 29.868 pixels and < 173 pixels.