A Multi-Resolution Approach to Point Cloud Registration without Control Points

Terrestrial photographic imagery combined with structure-from-motion (SfM) provides a relatively easy-to-implement method for monitoring environmental systems, even in remote and rough terrain. However, the collection of in-situ positioning data identification control points required georeferencing SfM processing is primary roadblock to using difficult-to-access locations; it also bottleneck time series. We describe novel, computationally efficient, semi-automated approach unreferenced point clouds (UPC) derived from terrestrial overlapping photos reference dataset (e.g., DEM or aerial cloud; hereafter RPC) order address this problem. The utilizes Discrete Global Grid System (DGGS), which allows us capitalize on easily collected information about camera deployment coarsely register UPC RPC. DGGS hierarchical set grids supports modified iterative closest algorithm natural correspondence between requires minimal interaction user-friendly interface, while allowing user adjustment parameters inspection results. illustrate two case studies: close-range (<1 km) vertical glacier calving front reconstructed cameras at Fountain Glacier, Nunavut long-range (>3 scene flat ice four overlooking Nàłùdäy (Lowell Glacier), Yukon, Canada. assessed accuracy by comparing RPC, as well surveyed points; consistency registration was difference successive registered surfaces roughly equal ground sampling distance consistent across steps. These results demonstrate promise acceptable accuracy, opening door new possibilities change-detection, such rates, surges, other seasonal changes field locations.