The Research Of Motion Blur Images Restoration

It is well-known that the images of moving objects, such as shooting moving vehicle or shooting by moving camera, will be blurred. The blurred image with a serious disturbance and inaccuracy, analytical processing or pattern recognition of blurred images with serious disturbance and inaccuracy is directly impossible to get accurate results, how to get accurate and useful information from the blurred images; and restore the information of blurred image are an important part of the intelligent image processing, it is also the basis and prerequisite for accuracy of subsequent processing. This paper mainly studies the restoration techniques of vehicle motion blurred images, the following work is completed in the paper:First, the article introduced restoration methods and development currently, discussed the typical problems in the recovery results, such as ringing and the difference of partial recovery, and analyzed existing approaches.Ringing seriously affects on the recovery results. Particularly, when using Wiener Filter recovers the blurred image. Among the experimental analysis, it is proved that not all the image restoration results by the Wiener filter emerge ringing, the restoration results of simulated blurred images are withnot ringing. Through the data analysis and the discussion on the differences between simulated blurred images and actual motion blurred images, the paper proposed a boundary-compensation method to suppress the ringing generation.The difference of partial recovery is that the restoration results is different for different areas. Considering the process of the blurred image, it is found that the blur parameters of blurred images are not identical, that is to say blur parameters are different during the different regions of the image. Further analysis the blur process, the article sorted out the relationship of blur length between the different regions of the blurred image, and proposed the new method of block-restoration, established principle of block-restoration according to the blur length of each block and the accuracy requirements, in order to obtain the consistent results.Fnally, deringing and blocked were applied to the actual vehicle blurred image, the recovery effect was consistent, and conquered the ringing effects, but the algorithm need to calculate the start and end of each block and the blur length of them, the intelligent of algorithm is difficulty, and the diversity of actual movement is the new challenge to the applicability of this algorithm, so the intelligent and applicability of the algorithm will is the main directions for future researches.