[4-1600-C] Postharvest/Food Technology and Process Engineering (4)

[4-1600-C-07] Myanmar Mango Maturity Prediction Based on Skin Color Using Machine Vision System

*RULIN CHEN¹, Dimas Firmanda Al Riza¹, Thwe Thwe Tun Naw², Phyu Phyu Leiyi², Aye Aye Thwe², Khin Thida Myint¹, Yuichi Ogawa¹, Tetsuhito Suzuki¹, Naoshi Kondo¹ (1. Kyoto University(Japan), 2. Yezin Agricultural University(Myanmar))

Keywords:Maturity prediction, Machine vision

Mango fruits mostly in Myanmar, are harvested without maturity grading, which will cause later problems in processing and transportation. Mangoes of different maturity level show difference in skin characteristics like color. As machine vision system can acquire visual information of an object and give feedback in a short time, it is likely to be an effective method to predict mangoes’ maturity based on their skin characteristics. In this study, to better evaluate mangoes’ skin characteristics, 3 color models (RGB, HSV, CIELAB) were compared and discussed. Sein ta Lone mangoes' skin characteristics during maturation were recorded as color images and fluorescence images using USB camera with white and UV LED illumination. Changes of color figures in 3 color models of 2 types of images during maturation were analyzed using Mat lab and compared. As a result, a* value from the CIELAB color model is observed an obvious change during mango maturation in color images compared to others.