Blind image super-resolution reconstruction based on PSF estimation

Blind image super-resolution reconstruction is one of the challenges in image restoration. In order to improve the spatial resolution of low resolution images, a blind image super-resolution reconstruction method is proposed. The point spread function (PSF) of the imaging system is estimated to approximate the low-resolution imaging model much more accurately. An error-parameter analysis method is proposed to estimate the parameters of Gaussian PSF. Utilizing Wiener filtering image restoration algorithm, multiple curves are generated at different parameters. By setting some thresholds, the size and deviation of the Gaussian PSF can be estimated automatically. The experimental results show that this algorithm achieves high precision, and validates the great importance of PSF estimation in image super resolution reconstruction. The experimental results demonstrate the effectiveness of the proposed algorithm.