Feature Extraction for Quality Modeling of Malang Oranges on an Automatic Fruit Sorting System

This paper evaluated some of the machine vision techniques to extract fruits characteristic, which in this case is Malang oranges. Appropriate algorithms are developed and implemented to extract features from local fruit based on Indonesian National Standard (SNI 3165: 2009) which is concerning to oranges. The research is done by analyzing fruit images, extracting HSV parameters and extracting feature using contour detection, hull convex and RGB histogram. A sensing machine which is consists of a photo box with a camera and a conveyor has been developed. The detection process can be done in real-time with the help of boxes equipped with adequate lighting. Convex hull analysis can be used to determine the diameter that has a great effect on the citrus fruit classification. While the red-green ratio can be used to label citrus fruits so that it can be used on a gradation-based fruit sorting machine. The performance was evaluated in terms of measurement accuracy which is above 88%. The research has the potential to be improved with the addition of an artificial intelligence-based decision system.