Since Martin Zimmermann's cinematographic reconstruction of the vascular anatomy of plants, there have been few systematic advances in our approach to this problem in spite of great progress in image processing and software in general. The main challenge in reconstruction of vascular anatomy in plants is really one of object recognition. First the cross-section of the stem, for example, needs to be located within the image; then vascular conduits need to be recognized within that cross-section. This falls under the rubric of image segmentation, the breaking up of an image into a number of meaningful sub-domains.