Determination of the atmospheric point spread function by a parameter search

Abstract

The result of blind deconvolution is a reconstructed image that has non positive intensities. The number of these non positive pixels has been used as an estimator for the departure from a perfect reconstruction. Simulations of reconstruction of objects convolved with a kernel having one or two parameters are shown to demonstrate the efficiency of the estimator. We thus present a technique of determining the unknown parameters of the point spread function by searching for the point in parameter space with the lowest number of nonpositive pixels. It is also shown that the parameters of the convolving kernel can be obtained even in the presence of noise. This method was validated using a long exposure image of NGC 1409.