Non-destructive NIRs-assessment of apple quality and ripening parameters, compared to conventional analysis by an appropriate statistical procedure
Zanella A.; Cecchinel M.; Cazzanelli P.; Coser M.; Panarese A. and Rossi O.
5th International Postharvest Symposium . Volume of Abstract . Verona, Italy 6-11 June 2004, p.122
2004
บทคัดย่อ
Non-destructive NIRs-assessment of apple quality and ripening parameters, compared to conventional analysis by an appropriate statistical procedure
The suitability of NIR-spectroscopy for the non-destructive quality assessment of apple fruits was investigated in the past 2 years. The measuring system (Sacmi F5,
Imola, Italy) is based on spectral analysis of NIR transmittance that is more accurate than the reflectance modus. The equipment is a mechanical set-up for commercial sorting-lines. The entire fruit is illuminated by an intensive halogen light source. The spectral changes of transmitted light contain information about chemical and physical composition, ripening and internal disorders. The aim of the study was to compare this novel instrument with traditional destructive techniques. In order to optimise the reliability of the NIR-measurements, specific calibration patterns for each apple variety were created. The cultivars under study were ‘Gala’, ‘Golden Delicious’, ‘Red Delicious’, ‘Jonagold’
‘Braeburn’, ‘Granny Smith’, ‘Fuji’ and ‘Cripps Pink’. Samples were characterised by a wide range of maturity stages. The detection of following parameters was evaluated: i) ripening, in terms of starch-index, endogenous ethylene and Streif-index; ii) quality, interms of firmness, total soluble solids and malic acid content; iii) internal disorders, like internal flesh browning, core flush and water core.
In order to compare the two different analytical methods, particular attention was given to develop an appropriate statistical procedure. Reliability and reproducibility were given by calculating intraclass correlation coefficients in addition to conventional values like correlation factor, SEC and SEP (standard error of calibration, prediction, resp.). Moreover, a graphic representation of the statistical comparison is given using the Bland-Altman (1999) plot.