Surface reconstruction of oil paintings for digital archiving

This paper proposes a method for reconstructing the surface shape of an oil painting for the purpose of digital archiving of art paintings. The painting surface is regarded as a 3-D rough surface with the dichromatic reflection property. First, imaging systems are shown for image acquisition and analysis. Next, an algorithm is presented for constructing surface height from the surface normal vectors estimated using high-resolution images. Moreover, we use the low-resolution surface height information of direct measurements from a laser scanning meter. The accuracy of constructed surface is then improved by incorporating the low-resolution height information to the high-resolution gradient images. Finally, all the estimates of the surface properties are combined for realistic image rendering. The feasibility of the method is shown in experiments using real oil paintings.