Validation of betacyanin content in Beta vulgaris leaves by using non-destructive measurement with Dualex™ leaf clip sensor or proximal sensing spectroscopy

Non-destructive measurements of leaf total betacyanin content (TBC) have been addressed for the first time by employing proximally sensed fluorescence, using Dualex™ leaf clip, and reflectance measurements, using a spectroradiometer, of Beta vulgaris leaves coupled with monovariate regression analysis and a multivariate signal processing framework. Both fluorescence-based and reflectance-based indexes have been investigated in terms of their capability of predicting TBC based on regression models, whereas a multivariate technique has been adopted to exploit the entire shape of leaf reflectance spectra, ranging from visible to short-wave infrared wavelengths. Results showed that the reflectance-based index outperformed the fluorescence-based index for TBC prediction and that usage of the entire reflectance spectra outperformed both of them, by providing much lower prediction errors (i.e., up to, on average, 148 μg∙g-1 fresh weight (FW) for TBC ranging in 100–1600 μg∙g-1 FW). Suitable multivariate signal pre-processing involving transformations of the reflectance spectra revealed to be beneficial to TBC prediction performance thanks to the enhancement of betacyanin-related spectral features. The multivariate framework also provides, as outcome, the subset of most relevant wavelengths for TBC prediction, which sets the basis for future developments of new TBC-related spectral indexes as well as design of low-cost devices for non-destructive leaf betacyanin assessment.