Analysis of the Effect of the Wax Coating on Firmness Prediction Model in Malang Apples Based on Visible and Near-Infrared (VNIR) Imaging

Abstract

Nowadays, wax-coated fruits to reduce respiration rate, maintain firmness, and postpone fruit ripening. Usually, the instrument that used to measure the quality of waxed fruit was destructive. The quality measurement of the fruits using Visible and Near-Infrared (VNIR) imaging has been studied. But, the quality measurement of the waxed fruits with VNIR imaging had never been done before. In this study, the effect of the wax coating on the firmness prediction model of Malang apples was further analyzed. The object that used was Rome Beauty variety of Malang apples. Beeswax was used to coat the Malang apples. Images were recorded using reflectance mode. Images were processed using several methods such as image correction, selection of Region of Interest (ROI), feature extraction, feature selection, and regression model. Regression Tree (RT) used as a feature selection and regression model algorithm. In this study, a regression model was built using non-waxed Malang apples. Next, the waxed Malang apples were tested with firmness prediction model of the non-waxed Malang apples. The evaluation parameters that used to evaluate the model were Root Mean Square Error (RMSE), determination coefficient $(\mathrm{R}^{2})$, and Residual Predictive Deviation (RPD). Prediction results for non-waxed Malang apples were 0.88 for $\mathrm{R}^{2}$, 6.65 for RMSE and 2.11 for RPD. Test results for waxed Malang apples to firmness prediction model of non-waxed Malang apples were 0.3 for $\mathrm{R}^{2}$, 16.8 for RMSE, and 0.85 for RPD. Based on these results, the wax coating on the surface of the fruits could disrupt the measurement results of VNIR imaging.