An algebraic approach to image restoration filter design

The image restoration problem in a linear imaging system is formulated as the problemof minimizing the effective radius of the system point spread function subject to a constraint on relative noise gain. The problem is solved by reducing the imaging system under consideration to an equivalent sampled system and subsequently optimizing the sampled system by algebraic techniques. The analysis applies to unsampled, line scanned, and sampled systems and explicitely accounts for spectrum foldover. Truncation problems are avoided in the discrete case by formulating the optimum processing array as the solution to a finite dimensional eigenvector equation.