Non-destructive methods for mango ripening prediction: Visible and near-infrared spectroscopy (visNIRS) and laser Doppler vibrometry (LDV)

With up to 19% of mango fruit being lost during ripening, the need for non-destructive technologies to predict internal physiochemical traits is paramount. This study compared two non-destructive technologies, visible and near-infrared spectroscopy (visNIRS) and laser Doppler vibrometry (LDV), for predicting the ripeness of mango fruit in two cultivars, 'Kent' and 'Keitt'. An internal quality index (IQI) in 'Kent' was predicted using visNIRS (RP2 = 0.729, RMSEP = 0.532) using partial least squares regression, which gave a single measure for ripeness incorporating firmness, sweetness, and pulp colour. This model was improved by using the sum of the individual sugar contents (glucose, sucrose, and fructose) over the conventional total soluble solids (TSS) measure. LDV provided poor predictions of firmness (R2 < 0.5) in both 'Kent' and 'Keitt' using least squares regression line. The resonant frequency, as measured by LDV, decreased linearly with time, while firmness quantified destructively (quasi-static) showed an exponential decrease, suggesting the vibrational and destructive firmness measure distinct characteristics, which would contribute to poor model performance. These results showed that LDV is not suitable for assessing mango ripening. While visNIRS models have been successful at predicting quality traits, our results suggested that using individual sugar content in place of TSS can improve the prediction of ripening. This understanding of the strengths and limitations of both visNIRS and LDV, and how they relate to destructive quality measurements, can be used to improve postharvest management practices whilst reducing commercial losses in the mango industry.