Prediction and Comparison of Models for Soluble Solids Content Determination in ‘Ya’ Pears Using Optical Properties and Diffuse Reflectance in 900–1700 nm Spectral Region

Abstract

The basic optical properties of fruit tissues are significant and helpful to be known when they are utilized for internal quality detection. The integrating sphere (IS) system was established to measure diffuse reflectance spectra (R') of peel-flesh layer, absorption coefficient (μ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">α</sub> ) and reduced scattering coefficient (μ' <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">s</sub> ) of flesh layer of `Ya' pear tissues using inverse adding-doubling (IAD) method in 900-1700 nm. Aiming at studying the relationship between soluble solids content (SSC) and optical properties (μ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">α</sub> and μ' <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">s</sub> ), partial least square (PLS) regression was employed to establish calibration models based on three wavelength regions. Results showed that better correlations were found using reflectance and μ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">α</sub> spectra than using scattering ones with full wavelengths in 900-1350 nm. Moreover, reflectance spectra (r <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</sub> = 0.8911, RMSEP = 0.4965 °Brix) with 27 effective wavelength (EWs) and μ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">α</sub> spectra (r <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</sub> = 0.8641, RMSEP = 0.5506 °Brix) with 14 EWs both performed satisfactory results using competitive adaptive reweighted sampling (CARS). Good result (r <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">p</sub> = 0.8856, RMSEP = 0.5081 °Brix) could also be found by the combination of reflectance and μ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">α</sub> spectra with 38 EWs. Results of independent validation test showed that EWs of μ <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">α</sub> spectra was more promising in pear SSC prediction even though the model stability was not as good as these obtained by reflectance. Therefore, more research needs to be done to improve the stability and accuracy of the IS system. This study illustrated the availability for multiple index quality attributes detection of fruit tissues by detecting reflectance and optical properties information using IS system.