ИСТИНА

The blurring of a two dimensional image is often modeled as convolution with a point spread function specific to the imaging instrument. In certain applications, (e.g. high energy radiography) standard techniques to reconstruct the point spread function that involve imaging a bright point source are not feasible. In this work, we will present an alternative method for point spread reconstruction that is suitable for applications where it is possible to image a sharp edge. The reconstruction is given as a solution to a Fredholm integral equation of the first kind. We model the problem stochastically and use Bayesian techniques for providing stable estimates of the solution as well as quantifying uncertainty in the estimate.