Keypoint-Based Stereophotoclinometry

This paper proposes the incorporation of techniques from stereophotoclinometry (SPC) in a structure-from-motion (SfM) system to estimate the surface normal and albedo at detected landmarks. In contrast to the traditional SPC paradigm, which relies on human-in-the-loop verification and a priori information to achieve accurate results, we forego the expensive maplet estimation step and instead leverage dense keypoint correspondences from a deep learning-based keypoint detection and matching method to provide the photogrammetric constraints. The proposed framework is validated on imagery of the Cornelia crater on Asteroid 4 Vesta.