Visual pose estimation and identification for satellite rendezvous operations

Satellites operating in close proximity require precise identification methods and pose estimation systems so that rendezvous and docking is possible. While dedicated RADAR and LIDAR systems have been applied to this problem, a purely visual method using one or more cameras provides a low-power and compact option that does not require dedicated ranging hardware. By means of visual feature detection and tracking across rapid, successive frames, features detected in two-dimensional images are matched and triangulated to provide three-dimensional feature maps using structure-from-motion techniques. Triangulated points are organized by means of orientation histogram descriptors and used to identify and track targets over time. The state variables with respect to the camera system are then filtered over periods of time to determine the relative movement of the spacecraft. This method is tested using laboratory images of nanosatellites with a simulated spacecraft movement model with varying parameters and conditions