File(s) not publicly available

Assessment of internal flesh browning in intact apple using visible-short wave near infrared spectroscopy

journal contribution
posted on 27.07.2018, 00:00 by Bed KhatiwadaBed Khatiwada, Phul SubediPhul Subedi, Clinton HayesClinton Hayes, LCC Carlos, Kerry WalshKerry Walsh
© 2016 Elsevier B.V.Certain cultivars of apple are prone to an internal flesh browning defect following extended controlled atmosphere storage. A number of (destructive) reference methods were assessed for scoring the severity of this defect in a fruit, including visual assessment, image analysis (% cross section area affected), International Commission on Illumination (CIE) chromameter Lab values of a cut surface and juice Abs420, of which visual scoring on a 5 point scale and a colour index based on CIE Lab were recommended. Non-invasive detection of this disorder using three instruments operating in the visible-shortwave near infrared (NIR) but varying in optical geometry (interactance, partial transmission and full transmission) was attempted. Quantitative prediction of defect level was best assessed using visible-shortwave NIRS in a transmission optical geometry, with a typical partial least squares (PLS) regression model with correlation coefficient of determination, R2p = 0.83 and root mean square of errors of prediction = 0.63 (5 point defect score scale). The binary classification approaches of linear discriminant analysis, PLS discriminant analysis, support vector machine approach and logistic regression were trialled for separation of acceptable fruit, with the best result achieved using the PLS discriminant analysis method, followed by linear discriminant analysis and support vector machine classification. Classification accuracy [(True Positive + True Negative)/(Positive + Negative)] on an independent validation population of >95% and a false discovery rate [False Positive/(True Positive + False Positive)]of <2% was achieved.

History

Volume

120

Start Page

103

End Page

111

Number of Pages

9

eISSN

1873-2356

ISSN

0925-5214

Publisher

Elsevier BV

Peer Reviewed

Yes

Open Access

No

Era Eligible

Yes

Journal

Postharvest Biology and Technology