Abstract
Near infrared spectroscopy (NIRS) has been widely used to assess horticultural fruits non-destructively for internal quality. However, it is far from perfect and new areas need to be explored for technological advancements. An optical property that has not been investigated in this field is the polarization of light. In this thesis, the feasibility of using polarized light, for non-destructive assessment of the internal quality of fruit, is studied.
A first experiment, involving 200 SunGoldTM kiwifruit, consists of a commercial NIRS instrument to generate either co-polarized or unpolarized light for internal fruit quality prediction of intact and peeled kiwifruit. Aquaphotomics and partial least squares regression (PLSR) were used to explain differences. Results suggested that more knowledge on how polarized light propagates through fruit was needed.
This led to the development of an NIRS polarimeter operating in the 600 - 1000 nm range, which was able to generate and measure polarized light in both transmission and reflectance. Using the Lu-Chipman decomposition method, polarization properties were obtained from experimentally measured Mueller matrices.
A transmission mode experiment generated Mueller matrices for skin peels and tissue slices from four different fruit types. Experimental Mueller matrices were decomposed to calculate the polarization properties of fruit tissue and skin peel separately. The depolarization coefficient dominated in all cases with fruit tissue slices being related to its scattering power. Skin peel varied considerably in order by sample type, possibly due to a combination of thickness and scattering power differences. A further transmission mode experiment measured the Mueller matrices with postharvest ripening of SunGoldTM kiwifruit. The depolarization coefficient of tissue slices held close relationships with the optical coefficients and firmness whereas correlations with the skin peel depolarization and firmness were much poorer. Linear retardance, possibly caused by a birefringent material such as cellulose, was flat with wavelength, and varied from sample to sample. The circular retardance of the tissue slices was effectively zero. However, the circular retardance of kiwifruit skin peels were statistically non-zero and trended down with increasing wavelength. It is suggested this may be due to the highly depolarizing nature of kiwifruit skin peel.
A reflectance study extended previous studies to answer the question: Can polarized light be used to non-destructively assess internal fruit quality? The degree of polarization (DOP) of the reflected light were measured for two apple cultivars, ’Royal Gala’ and ’Bay Queen’. Results show a decline of DOP with fruit age but very poor relationships with firmness measures, which suggested that the DOP is likely to be used for fruit firmness measurements.
This study has advanced our knowledge of NIRS technology on horticultural fruit using polarized light. The feasibility of using polarized light seems unlikely for the purpose of direct non-destructive assessment of internal fruit quality, whether for off-line measurements or for online grading and sorting due to the depolarizing fruit skin. Future applications of NIRS polarized light on fruit is likely to be limited to excised thin samples, skin and/or only the very near surface regions of intact fruit.