A. Peirs, K. Touchant, A. Schenk, B. Nicolaï
Determination of the optimal harvest date of apples is very important. Fruits harvested too early have inferior organoleptic quality and are prone to physiological disorders such as superficial scald, whereas fruits harvested too late develop a soft and mealy texture during cold storage. The ideal picking period is determined based on the time course of individual or combinations of chemical and physical properties during a well-defined period before harvest. In this paper, the use of FT-NIR spectroscopy has been examined to reduce the substantial amount of experimental work required to predict the optimal harvest date.
NIR-spectra were used to create PLS (partial least square) models that predict quality attributes. Fruit maturity was very well correlated with the reflectance spectra. Because of cultivar differences, validation correlations were between 0.94 and 0.97. The validation correlation of models for the Streif-index was between 0.87 and 0.95, depending on the cultivar. Models for correlating the spectra with fruit respiration had a validation correlation ranging between 0.87 and 0.94. Validation correlation of the models for the prediction of the respiratory minimum ranged between 0.96 and 0.97, which means a RMSEP between 4.2 and 6.9 days. This is very precise, taking into account the large variation in the apples on one tree.
Peirs, A., Touchant, K., Schenk, A. and Nicolaï, B. (2001). FT-NIR SPECTROSCOPY TO EVALUATE PICKING DATE OF APPLES. Acta Hortic. 553, 477-480
DOI: 10.17660/ActaHortic.2001.553.112
Picking date, apple, ripening, PLS, prediction

Acta Horticulturae