Image analysis of changes in surface color of chocolate

Abstract

Blooming or the migration of fat to the surface of chocolate results in color changes and development of non-uniform color patterns. These phenomena were assessed during storage of milk chocolate tablets (cycling temp. between 16 and 28 °C for 52 days) by a computer vision system and image analysis. Eight features were extracted from images ( L*, a* and b* values, whiteness index, chroma, hue, % bloom and energy of Fourier). Major changes occurred after day 36 of storage, coincidental with visual perception. Initially, white specks emerged on the brown background but were superseded by the development of a whitish color extending over most of the surface. L*, whiteness index, a* and chroma correlated well with values taken with a commercial colorimeter ( R 2 > 0.70). Changes in image texture (energy of Fourier) followed a similar trend as color changes. The sequential forward selection strategy allowed correct classification of 97.8% of samples into four classes with only five features. The computer vision system has the capability to quantify overall changes as well as particular features over the whole chocolate surface thus enabling customization and standardization for quality assessment.