Color measurement: comparison of colorimeter vs. computer vision system

Abstract

The aim of this work was to compare two food color measurement techniques, the traditional tristimulus colorimeter and an image analysis system. In this sense a computer vision system was developed, consisting of a digital camera, a controlled illumination environment, and a software package to process the images. The conversion between color spaces was performed employing empirical mathematical models; a standard color chart was used for its calibration. The color of 40 samples of raw and processed foods was measured in the CIELAB color space with the computer vision system and a colorimeter. The equivalence between both techniques, for individual L*, a* and b* values, was determined using appropriate hypothesis tests. For most samples both systems provide equivalent results, although the total color difference ∆Ε was high enough to be noticeable. The average ∆Ε was 5.88 ± 3.32, with an average absolute ∆L* = 2.79 ± 2.42, ∆a* = 3.02 ± 2.94; ∆b* = 2.84 ± 2.53. In addition, the color measured by the image analyses technique seemed to be more similar to the real ones.