Hierarchical Image Processing for CSO Resolution

Observation of closely-spaced objects using limited-resolution infrared sensor systems can result in merged measurements, which can hamper the performance of both tracking algorithms and discrimination. Algorithms intended to improve image resolution capability are often processor-intensive. This can limit their applicability for space systems onboard processing, particularly in the short timelines associated with kinetic interceptors. We propose to develop an algorithm for resolution of closely-spaced objects using a hierarchy of spatial filters, deconvolution, and constructive particle distribution with Expectation Maximization (EM) clustering. With this hierarchy, we expect to provide improved object resolution while conforming to space system processing time constraints.