Kai-June Xu¹*, Pei-Li Sun¹*, Raymond Chiang² and Fu-Ling Chan³
¹ National Taiwan University of Science and Technology, Taipei City, Taiwan
² Central Engraving and Printing Plant, New Taipei City, Taiwan
³ Taiwan Textile Research Institute, New Taipei City, Taiwan
*Corresponding author: Pei-Li Sun, plsun@mail.ntust.edu.tw
Keywords: Color measurement, mobile phone camera, color rendering chart, lighting environment, color
correction, color metamerism
Poster presentation video: https://youtu.be/ACVn8QSKi0A
Camera is a useful tool for color measurement of flat surface. However, there are many factors affects the accuracy of the measurement such as illuminant metamerism, sensor metamerism, geometry metamerism, color enhancement of the camera and optical property of the surface. In this study, the effects of 5 factors including camera choice, substrate, light source, regression method and sample placement method were estimated. 3 cameras (including 1 DSLR and 2 mobile-phone cameras), 3 types of flat samples (matte paper, glossy paper and plastic chips in 3 mm depth), 4 light sources (D65/CWF/VVF/TL84), 3 color correction methods and 2 placement methods were tested. 24 color samples were photographed with a ColorChecker Proof attached on top of them. The experimental results show that the errors of direct color space conversion are quite large, with an average of more than 10 ∆E00. Using the color chart as a reference to assist color measurement can reduce the color differences to a half. The best color correction method is "color difference-weighted high-order linear regression". Due to the translucent nature of plastic color samples, it is necessary to compensate for "light leakage" in the XYZ color space to reduce color prediction errors. In terms of the effects of the camera, light source, substrate, and placement method, the results show that the performance of the cameras varies, but the differences are not big. Under the four tested light sources, D65 and CWF perform better, while TL84 performs the worst. Among the three type of substrate, colors of inkjet printing on matte paper are most predictable and stable, whereas colors of plastic samples are more unpredictable. For placing the plastic samples, lying flat mode is better than wall- mounting mode in the light booth. The results provide valuable guidelines for improving the accuracy of camera-based color measurement.
Comments