Remote sensing image restoration using estimated point spread function

In order to reduce image blur caused by the degradation phenomenon in the imaging process, the acquired images of the space remote sensing camera are restored. First, the frequency-domain notch filter is adopted to remove strip noises in the images. Then degradation function, which is referred to as the point spread function (PSF) of the imaging system is estimated using the knife-edge method. To improve the accuracy of the estimation, the estimated PSF is adjusted with Gaussian fitting. Finally, the images are restored by Wiener filtering with the fitted PSF. The restoration results of the remote sensing images show that almost all strip noises are eliminated by the notch filter. After denoising and restoration, the variance of the remote sensing image worked with in this paper increases 30.979 and the gray mean gradient increases 3.312. Due to Gaussian fitting, the accuracy of the PSF estimation is heightened. Image restoration with the final PSF is benefit to interpreting and analyzing the remote sensing images. After restoration, the contrasts of the restored images are increased and the visual effects become clearer.