Prediction of the optimal picking date of different apple cultivars by means of VIS/NIR-spectroscopy

The use of visible/near infrared (VIS/NIR) spectroscopy was evaluated to de termine the internal quality and the optimal harvest dates of apples non-de structively. Calibration models were constructed with data from eight culti vars, three orchards and 2 years, in order to make the models as robust as possible for future use. The prediction of the maturity, defined as the num ber of days before commercial harvest, was reasonably accurate. The most ro bust model predicted the maturity with a validation correlation of 0.90 (SE P = 7.4 days). The prediction of maturity, according to the Streif index, s howed a validation correlation of 0.84 (SEP = 0.18 kg/% brix x starch index ) for one orchard. Maturity was orchard-dependent, however, and as a conseq uence, a combined prediction equation was not accurate. Individual quality characteristics (soluble solids, Streif index, acidity and firmness) were w ell predicted. The calibration model for soluble solids content resulted in a validation correlation of 0.84 (SEP = 0.73% brix) for the results over 2 years from one orchard, but like the Streif index, was orchard-dependent a nd appeared to account largely for the orchard dependence of the latter. Ac idity and firmness were predicted with a validation correlation of 0.80 and 0.78 and SEPs of 2.07 ml NaOH and 1.13 kg for, respectively, two and three orchards over the 2 years. (C) 2001 Elsevier Science B.V. All rights reser ved.