Handheld Short-Wavelength NIR Spectroscopy for Rapid Determination of Sugars and Carbohydrate in Fresh Juice with Sampling Error Profile Analysis
Wanchao Chen,Hui Li,Feiyu Zhang,Weimin Xiao,Ruoqiu Zhang,Zengkai Chen,Yiping Du
DOI: https://doi.org/10.1016/j.infrared.2021.103732
IF: 2.997
2021-01-01
Infrared Physics & Technology
Abstract:Fresh fruit and vegetable juices have become fashionable drinks, in which sugars and carbohydrates contents have always been focused by consumers. Herein, a rapid and convenient analytical method based on handheld short-wavelength NIR (SW-NIR) spectroscopy and PLS regression modeling was developed. For accurate results, the strategy of Sampling Errors Profile Analysis (SEPA) was applied to make the results more accurate. For the determination of sugars content, the obtained the interquartile ranges (IQR), R-p(2), RMSEP and the ratio of performance to deviation (RPD) were 0.30%, 0.89, 0.95% and 3.06, respectively when raw spectra were used for modeling. For carbohydrate, the corresponding indicators were 0.14% 0.96, 0.63% and 5.06, respectively if the first derivative was used to pretreat the spectra. Compared with the modeling results from the desktop short-wavelength NIR devices, the RMSEP and RPD was reduced by 16.8, 20.3% of sugar and 22.2, 28.7% for carbohydrate, respectively. The current results demonstrated that the handheld SW-NIR sensor enables the development of a broad range of sensing integrated applications on determine sugars and carbohydrate in fresh juice.