Photometric Stereopsis for 3D Reconstruction of Space Objects

The use of photometric stereopsis approaches to estimate the geometry of a resident space object (RSO) from image data is detailed. The set of algorithms and methods for shape estimation form an integral element of a Dynamic Data Driven Application System (DDDAS) for enhancing space situational awareness, where, sensor tasking and scheduling operations are carried out based upon the RSO orbital and geometric attributes, as estimated from terrestrial and space-based sensor systems. Techniques for estimating the relative motion between successive frames using image features are used for data alignment before surface normal estimation. Mathematical models of photometry and imaging physics are exploited to infer the surface normals from images of the target object under varied illumination conditions. Synthetic images generated from physics based ray-tracing engine are used to demonstrate the utility of the proposed algorithms.The proposed framework results in a estimates of the surface shape of the target object, which can subsequently used in forward models for prediction, data assimilation and subsequent sensor tasking operations. Sensitivity analysis is used to quantify the uncertainty of reconstructed surface.