Integration of Hyperspectral Reflectance and Fluorescence Imaging for Assessing Apple Maturity

Abstract

Fluorescence and reflectance are two different forms of light interaction with matter, and they can be complementary in measuring fruit quality and condition. The objective of this research was to develop an integrated hyperspectral reflectance and fluorescence imaging technique for measuring apple maturity. Both fluorescence and reflectance scattering images were acquired using a hyperspectral imaging system covering the wavelengths of 500 to 1000 nm from 'Golden Delicious' apples harvested over a 4-week period. Standard destructive tests were performed to measure multiple maturity parameters (flesh and skin color, firmness, soluble solids, starch, and titratable acid). The spectral fluorescence and reflectance scattering profiles were described by a two-parameter Lorentzian function. Multi-linear regression prediction models were developed relating Lorentzian parameters to individual maturity parameters for each sensing mode and their combined data. The fluorescence prediction models had consistently lower correlations with individual maturity parameters than the reflectance models. The integration of reflectance and fluorescence improved maturity measurements over either reflectance or fluorescence; the improvements in correlation were noticeable for most parameters (up to 12% for titratable acid). Since fluorescence and reflectance measurements were made with the same imaging system, the integrated technique can provide better assessment of apple fruit maturity and quality.