Photogrammetry and knowledge representation in computer vision

Photogrammetry and computer vision are closely related disciplines with differing goals. This paper explores the ramifications of those differences with an eye toward identifying opportunities for leveraging the contributions of both fields. We identify some of the thorny issues that computer vision researchers are wrestling with and point out some of the areas where advances in photogrammetry can help. In most cases, further progress will require a shift from the conventional techniques of photogrammetry to ones that are more compatible with the ground-level, real time, and fully automated constraints that are emphasized in computer vision.