Colour is a powerful communication element in most forms of cultural heritage. This importance of colour notwithstanding, the documentation of cultural heritage typically captures the geometrical aspects and seldom the spectral dimensions of an artefact. This is partly because the science of colour (called colorimetry) is non-trivial. In addition, capturing accurate colour data with digital cameras remains challenging due to the operating principle of standard imaging sensors and the need for a stable and well-characterised illumination source. Despite these limitations, the heritage science project INDIGO made it one of its central aims to generate colour-accurate photos from graffiti captured with standard digital cameras in varying outdoor illumination conditions. This paper first discusses the importance of colour accuracy in graffiti documentation. Afterwards, the text details (in a non-mathematical manner) essential colorimetric and camera principles that underlie the generation of colour images from raw image sensor data. This in-depth coverage supports clarifying the main hurdles to accurate photo colours. Finally, the paper introduces the open-source COOLPI software resulting from this research. We are confident that COOLPI will benefit any other heritage documentation project, or any application where digital cameras play a fundamental role in acquiring correct colour values.