Achievability of Cramer-Rao Lower Bounds by Multi-Frame Blind Deconvolution Algorithms, Part 2: PSF Estimation
AIR FORCE RESEARCH LAB KIRTLAND AFB NM DIRECTED ENERGY DIRECTORATE
Pagination or Media Count:
Cramer-Rao lower bound CRB theory can be used to calculate an algorithm-independent lower bound to the variance of any unbiased estimate of an unknown parameter. The theory also applies to joint estimation of multiple unknown parameters, to functions of estimates, and to estimates that have known bias gradients. CRBs are guaranteed to be lower bounds, but may not be achievable in practice. In particular, it is well known that algorithms that minimize cost functions to generate estimates have difficulty in achieving the CRBs for low signal-to-noise ratios. Our interest is in the achievability of the CRBs by algorithms that use a multi-frame blind deconvolution MFBD framework. In previous work, we analyzed the achievability of CRBs for MFBD-based estimation of object energy spectra. Here, we present initial results from our extension of this previous work to the analysis of the achievability of CRBs for MFBD-based estimation of point spread function PSF energy spectra that are estimated jointly with the object energy spectra.
- Numerical Mathematics