3D digital reconstruction of conglomerate: A comparative dataset evaluating micro-CT versus optical serial grinding with color-based segmentation

This data article presents two complementary three-dimensional (3D) image datasets of three quartzite conglomerate samples, acquired via X-ray micro-computed tomography (micro-CT) and serial grinding with optical scanning, respectively. The micro-CT data demonstrate the limitation of X-ray attenuation-based segmentation for samples where gravel clasts and matrix exhibit similar X-ray absorption, preventing reliable delineation of boundaries. In contrast, the serial grinding dataset provides high-contrast color images of successive surfaces. By splitting these images into red, green, and blue channels, the edges between clasts and matrix are significantly enhanced, enabling effective segmentation via a marker-controlled watershed algorithm. The dataset includes raw, aligned, channel-split, and segmented files for each sample, offering a direct comparison between the two imaging techniques. This data is a resource for researchers in visual paleontology, petrology, and porous media analysis who work with materials lacking X-ray contrast. It also illustrates a complete workflow from optical tomography acquisition to digital 3D reconstruction and highlights both the utility and the z-axis resolution limitation of the manual Croft parallel grinder.