A WORKFLOW FOR GEOMETRIC COLOUR PHOTOGRAPHY OF PAINTED SURFACES

Abstract. Colour fidelity is vital when documenting painted surfaces. The 2.5D nature of many painted surfaces makes orthophotos and digital surface models (DSMs) common products of the documentation process. This paper presents a workflow to combine photographic and photogrammetric methods to produce aligned colour and depth (orthophotos and DSMs). First, two photogrammetric software (Agisoft Photoscan and Capturing Reality Reality Capture) were tested to determine if they adjusted the colour data during the processing stages. It was found that Photoscan can produce 16-bit orthophotos without manipulating the data; however, Reality Capture is currently limited to 8-bit results. When capturing a surface using photogrammetry, it is common to use the same data for colour and depth. The presented workflow, however, argues that better colour accuracy can be achieved by capturing the two datasets separately and combining them in photogrammetric software. The workflow is demonstrated through the documentation of an unnamed religious painting from the 17th century.