As high-resolution conservation images, acquired using various imaging modalities, become more widely available, it is
increasingly important to achieve accurate registration between the images. Accurate registration allows information
unavailable in any one image to be compiled from several images and then used to provide a better understanding of
how a painting was constructed. We have developed an algorithm that solves several important conservation problems:
1) registration and mosaicking of multiple X-ray films, ultraviolet images, and infrared reflectograms to a color
reference image at high spatial-resolution (200 to 500 dpi) of paintings (both panel and canvas) and of works on paper,
2) registration of the images within visible and infrared multispectral reflectance and luminescence image cubes, and 3)
mosaicking of hyperspectral image cubes (400 to 2500 nm).
The registration/mosaicking algorithm corrects for several kinds of distortion, small rotation and scale errors, and
keystone effects between the images. Thus images acquired with different cameras, illumination, and geometries can be
registered/mosaicked. This automatic algorithm for registering/mosaicking multimodal conservation images is expected
to be a valuable tool for conservators attempting to answer questions regarding the creation and preservation history of
paintings. For example, an analysis of the reflectance spectra obtained from the sub-pixel registered multispectral image
cubes can be used to separate, map, and identify artist materials in situ. And, by comparing the corresponding images in
the X-ray, visible, and infrared regions, conservators can obtain a deeper understanding of compositional changes.