Improved Methods of Standardization for Color Measurements of Processed Tomatoes

This thesis presents the results from a study of optical properties of tomato paste. These optical properties were based upon diffuse reflectance measurements taken using a spectrophotometer and a specially designed hyperspectral imaging system. The scatter coefficient was estimated using two methods. The first method was based on the diffuse thickness measurements using a HunterLab spectrophotometer. The second method used diffuse reflectance values measured with a hyperspectral camera. A diffuse reflectance model developed by Farrell et al. (1992) was used to estimate the reduced scatter coefficient using non-linear regression applied to the hyperspectral image data. The diffuse reflectance model of Farrell et al. (1992) fit the collected data well (R 2 for all the samples was between 0.98 to 0.91) and the estimation of reduced scatter coefficient by the diffuse reflectance data were better (more theoretically robust) than the estimation by diffuse thickness measurement data.;The effects of the different break types, screen sizes and particle size of the tomato paste samples on the scatter coefficient was also studied. Both hot break and cold break tomato paste samples were studied, and there were six paste finish textures (expressed as process screen sizes 0.020", 0.033", 0.045", 0.060", 0.156" and 0.250") used for the experiment.;The second part of the thesis focused on the design of a shelf stable, light scattering, color glass reference standard for use as a color calibration hitch standard and its performance in comparison with the current standard used by the industry. The current standard used in the tomato industry is an opaque red tile developed by HunterLab which does not take into account the scattering of light caused by the translucency of tomato paste. The new light scattering color glass reference standard was made with three parts: a light scattering filter on the top to account for the scattering of light that occurs in the tomato paste, a red glass filter in the center to mimic the color of the tomato paste and a white Teflon disc at the bottom to reflect light and help in matching the actual reflectance level of reconstituted tomato paste. The reflectance values of this reference standard are similar to that of 8.5% NTSS paste, and also mimicked the scattering properties of tomato paste better than the Hunter red tile. Five hitching standards were examined using two different spectrophotometers to study the effectiveness of the new light scattering glass standard in reducing the inter-instrument agreement error. The delta a:b ratio was calculated for each of the hitching standards. This reading was the difference in the a:b ratio measured from the master spectrophotometer and the a:b ratio measured from the two spectrophotometers. The mean delta a:b ratio for the new light scattering color glass reference standard was 0.004. The diffuse reflectance spectra and the scattering properties of the new light scattering color glass reference standard closely match the optical properties of the actual tomato paste than the red tile, but the study was not able to show a difference between the performance of the new light scattering color glass reference standard and the hunter red tile as hitching standards for the two spectrophotometers used for the experiments.