| In many cases, specifically space related, it is difficult to match an object point directly to its corresponding image point in order to determine pose. To avoid this requirement, object points can instead be matched to corresponding areas in a given image and using convex Linear Matrix Inequalities (LMI) pose estimation (PE) can still be achieved. With the increasing availability of imaging platforms, due to more Nano-Satellites and Unmanned Aerial Vehicles being deployed, space situational awareness research has become an increasingly important topic. Processing data from these multiple sources and views could greatly improve the pose estimation algorithms which currently utilize only a single image.;In this thesis, the PE method based on points to regions correspondence is extended to utilize multiple camera views. The effect of multiple viewpoints is verified by simulation results, along with an analysis of convex region shape and placement. The simulation program, Satellite Toolkit (STK), was used to create scenarios in which a specific object (in this case a known 3D satellite) was tracked via smaller camera satellites. |
