COLOR-BASED ANALYSIS FOR NON-DESTRUCTIVE QUALITY EVALUATION OF SIAMESE ORANGE (Citrus nobilis) DURING STORAGE IN ROOM AND COLD TEMPERATURE

Abstract

This research aimed to develop image processing program to extract several features of Siamese orange images and to determine the appropriate key features of the orange images which were highly correlated with physical and chemical characteristics of Siamese orange. A sample size of 210 oranges was stored in a room temperature (25-27 °C) and cold temperature (8-10 °C) for 30 days. All oranges were captured their images using a webcam and were subsequently measured their physical and chemical characteristics, i.e. weight, hardness, and total soluble solids (TSS). The visual parameters of images measured were binary area in pixels, RGB color, HSV color, CIE-Lab color and gray value. The results showed that there was a high correlation between fruit weight and binary area in both room and cold temperature storage with r2= 0.8197 and r2= 0.8291, respectively. On the other hand, color values cannot be used to estimate fruit hardness since the correlation coefficient was too small. The highest correlation coefficient between them was r= 0.114 which was achieved from the correlation between fruit hardness and hue value. The r value of TSS and some color components was nevertheless relatively strong. Based on the experiments, hue value of HSV color model and a* component of CIE-Lab color model have fairly strong correlations with TSS of oranges stored in room temperature which were indicated from the lvalue of 0.7473 and 0.7029,respectively.