Extraction of Colour Information from Digital Images Towards Cultural Heritage Characterisation Applications

Abstract

Colour is a primary characteristic of visual perception and a very important characteristic to describe cultural heritage works in a variety of shapes, sizes and materials. To quantify colours, portable colorimeters are often carried and used for measuring colour of interest. Spatial resolution of colorimeters is manufacturer and model dependent and typically worse than 3 - 10 mm in diameter. The colour measurement has to be done on site at the time of interest. It poses very significant limitations in colorimetric characterization of cultural heritages. In this paper, the possibility of extraction of colour information from digital photographs, scanned images and video files using customized image processing/analysis software (PicMan) was investigated for cultural heritage characterization applications. As colour information extraction examples, a commercial Gouache paint colour card, a digital photograph, a portrait of a Korean scholar of Joseon Dynasty and a severely deteriorated old medical text book before and after restoration were examined.  Colour information from various digital images was successfully extracted from points and regions of interest in RGB, HSV, L*a*b*, Munsell colour and hexadecimal colour code formats. The spatial resolution of colour information extraction is as small as a single pixel in a digital image.