Application of Effective Wavelength Selection Methods to Determine Total Acidity of Navel Orange

Yong Hao,Xudong Sun,Hailiang Zhang,Yande Liu
DOI: https://doi.org/10.1166/sl.2011.1385
2011-01-01
Sensor Letters
Abstract:Three effective wavelengths (EWs) selection methods combined with visible and near infrared (Vis/NIR) spectroscopy were investigated to determine the total acidity (TA) of navel orange, including successive projections algorithm (SPA), monte carlo uninformative variables elimination (MCUVE) and randomization test (RT). Partial least squares (PLS) regression was used in the analysis of Vis/NIR spectroscopy, and three calibration models using Ews selected by SPA, MC-UVE and RI were developed. The results indicated that both three EWs selection methods can build a parsimonious model, and MCUVE-PLS achieved the optimal performance for TA prediction, the root mean square error of prediction (RMSEP) and calibration variables were reduced from 0.042 and 1451 to 0.035 and 710, and the correlation coefficient (r) was improved from 0.926 to 0.954. The overall results indicated that Vis/NIR spectroscopy incorporated to MCUVE-PLS could be applied as an alternative fast and accurate method for the determination of TA in navel oranges.
What problem does this paper attempt to address?