Calibrating Total Acidity of Tomato Products Using Ga and Vis/Nir Spectroscopy

Jiansong Yao,Haiqing Yang,Yong He
2009-01-01
Abstract:Non destructive optical methods based on visible and near infrared reflectance spectroscopy(Vis/NIRS) have been used for estimation of physiological properties of batches of fruit and vegetable products. The objective of this study was to evaluate the application of Vis/NIRS in measuring total acidity of tomato. Reflectance spectra in the 360-1050nm range were scanned on 200 tomato samples which were separated randomly into two groups: 150 for calibration set and other 50 for prediction. In order to examine the capability of using multiple wavelengths for prediction modeling, a genetic algorithm was adopted for the optimal selection of them. Three best wavelengths are 506.04nm, 667.50nm, and 837.23nm. The correlation coefficient of the model built on calibration set is 0.82675, which is better than one created at single wavelength. In prediction set, correlation coefficient of the model reaches 0.89789. These results suggest that Vis/NIRS measurements in effective wavelengths could be used to assess the internal quality of certain tomatoes, which can support further study on the design and implement of measuring meter of internal quality of tomato.
What problem does this paper attempt to address?