Research Of Motion-blurred Images Restoration Technology

In the field of digital image, image restoration is an important issue and also the difficulty. The purpose of image restoration is to improve the blur image quality, so that the gap between it and the original image is minimized, its recovery process through the establishment of a mathematical model of image degradation and then recover the original image by filtering. Motion blurred image restoration in image restoration is a very common type and also is very difficult problem.The reason of that is in the process of acquiring the image because of the relative displacement between object and camera equipment, Its recovery process is to simplify the movement patterns and identify image degradation model, then reverse the degradation process is derived to obtain the restored image. Motion blurred image restoration is currently widely used in aerospace, traffic monitoring, medical, forensic and other areas, Therefore, the motion blur image restoration study has important practical significance. Full text main work is as follows:(1) Introduces the research background and significance of the image restoration techniques and some basic knowledge about image restoration, followed by an analysis of the main sources of noise, characteristics and classification, and describes the image degradation model, especially under the uniform linear motion based on motion blurred image degradation model, And elaborated on its restoration works. Then introduced the image restoration quality assessment in two ways: a subjective evaluation and objective evaluation.(2) In determining the point spread function, a detailed analysis of the use of radon transform strike fuzzy angle and proposed improvements, allows to determine the angle of the blur more accurate. Use radon transform and differential autocorrelation were fuzzy scale strike, and these two methods were compared, analyzed the advantages and disadvantages between them.(3) Restoration algorithm introduces three kinds of the motion blurred image, and describes their advantages and disadvantages of these types of algorithms are analyzed, and on the basis of previous to the Wiener filter has been improved, for Wiener filter in the K value determines improved, improved Wiener filter can accurately determine the K value at the same time, effectively reduce the amount of calculation and the search range and can be locked in K values within a very small range.(4) Using MATLAB software and improved Wiener filter for motion blurred image to do restoration processing.In summary, after this paper do motion blurred image processing, and through the restored image and the experimental data we can know that: The proposed algorithm is superior to the ordinary Wiener filtering algorithm, when under uniform linear motion blurred image recovery process can achieve better results.