Modelling of Physical Properties for Prediction of Internal Quality Attributes of Kinnow

Abstract

Kinnow mandarin is a common citrus fruit grown commercially in north India. It is a rich source of vitamin C having high juice content and relished for its distinct flavour. Outer skin blemishes and shiny appearance are the primary parameters for quality assessment after harvesting, but the internal quality attributes of fruit are ignored due to lack of sorting methods. Models were developed for prediction of total soluble solids (TSS in ^oBrix) and Brix to acid ratio (BAR) from the physical attributes such as geometrical mean diameter (Dg), colour (Hunter a/b ratio)
and fruit density (Ï) measured non-destructively. The physical and chemical attributes were fitted in eight models using multiple linear regressions (MLR). The correlation coefficients for calibration and prediction of model (no. 4) fitted for variables Dg, colour and Ï for determination of TSS were 0.86 and 0.85, respectively. Similarly, the model (no. 8) relating three physical variables with BAR had correlation coefficients for calibration and prediction of 0.83 and 0.81, respectively. Besides satisfactory correlation coefficients, both the models exhibited low standard error of calibration and prediction as 0.95, 0.98 and 1.88, 1.94, respectively. The modelled correlation between physical and
internal quality attributes can be a basis for development of low cost non-destructive sorting of kinnow mandarin.