Application of spectra pre-treatments on firmness assessment of intact sapodilla using vis-nir spectroscopy

Abstract

This study aimed to obtain the best calibration model from various spectra pre-treatment methods to assess sapodilla fruit firmness using vis-nir spectroscopy. Before the spectra data measurement, samples were treated with storage of 0, 5 and 10 days at room temperature. Spectra data measurement was carried out using the NirVana AG410 visible and near infrared spectrometer from 312 to 1050 nm with interval of 3 nm. RAW spectra were pre-treated using the multiplicative scatter correction (MSC), standard normal variate (SNV), and Savitzky-Golay first derivative (dg1) with 9 points of smoothing. The calibration model was developed using PLS (partial least squares) method. Validation was done by K fold cross validation method. The results showed the MSC and SNV spectra were able to eliminate noises of RAW spectra, whereas in the dg1 spectra, noises were still visible. The best model was acquired by SNV spectra with R2 (coefficient of determination) of calibration and validation of 0.882 and 0.870, root mean square error of calibration (RMSEC) and root mean square error of cross validation (RMSECV) values of 2.92 and 3.08, and the ratio of performance to deviation (RPD) of 2.76. The result indicated the spectra pre-treatments were able to improve the accuracy of calibration model on assessment of sapodilla fruit firmness.