Spectral method for estimating the point-spread function in the task of eliminating image distortions
Valery S. Sizikov · Journal of Optical Technology · 2017
This paper gives a further development of a method for estimating the parameters of the point-spread function (PSF), based on the Fourier spectrum of a distorted image. This method makes it possible to estimate the PSF parameters: the angle θ and magnitude Δ of the image smearing, as well as the size ρ or σ of the image-defocusing spot. This is important for enhancing the image-reconstruction accuracy. The spectrum of a smeared image is compressed in the smearing direction, and this makes it possible to estimate θ and Δ. The spectrum of a defocused image is also compressed more strongly, the larger the defocusing spot. New, more accurate estimates of the defocusing parameters are derived, ρ using the Bessel function and σ using the three-sigma rule, and the smearing parameters θ and Δ are estimated using the Nyquist frequency. Numerical examples of the use of this technique are given. The technique thus developed can be used to enhance the accuracy with which smeared and defocused images are restored by mathematically processing them.