Techniques for characterizing tow architectures and defects in woven ceramic composites are required for generating high-fidelity geometric models and subsequently probing effects of defects on composite performance. Although X-ray computed tomography (CT) has been shown to provide the requisite information with potentially sub-m resolution, the technique is inherently limited to probing only small volumes: on the order of a few unit cells of typical weaves. Here, we present an assessment of the efficacy of a complementary 2D technique, based on surface topography mapping via 3-D (three-dimensional) digital image correlation (DIC), with potential for ascertaining long-range features in weaves and defects that cannot be gleaned from CT imaging alone. Upon comparing surfaces reconstructed from CT and DIC data, we find that DIC is capable of resolving surface heights with a root mean square(RMS) error of similar to 10m (about twice the CT voxel size, 4.4m) and a spatial resolution of similar to 20m over areas of several cm(2). Achieving this level of resolution requires use of sufficiently small speckles (similar to 50m) and small subset size (similar to 300m) relative to the characteristic tow dimensions (similar to 1mm). The error is somewhat higher (about 20m) in areas where surface discontinuities or rapid changes in topography exist (e.g., at tow boundaries).