Identification of motion blur direction based on analysis of intentional restoration errors

In this paper, a novel algorithm is proposed to estimate the direction of motion blur given only the blurred and noisy image. The identification approach is based on the analysis of restoration error characteristics due to the intentional use of a ramp PSF along horizontal direction. Strong strips along the direction of true blur appear in the restored image can be strengthened in vertical derivative of the image. Radon transform is incorporated to locate the blur direction accurately on the Fourier transform of the image derivative. Experiments on simulated images and real-world images show the power of our algorithm and its robustness to noise. Satisfactory results are still given even when the noise level is up to 10 dB.