A New Restoration Algorithm Of Digital Video

With the rapid development and progress of digital technology, its application scope has been wider and wider, as an important branch in digital technology, digital image technology has been widely applied in real life in many aspects. Image is an important method to get information from the outside world to us, and at the same time, it records a lot of precious historical material and film material with its special form, these film materials are most stored in film, because of many factors, there are different damages in these film materials, mainly including brightness flicker, shaking, small spots, scratches and some large patches. These film materials are so precious that their protection and restoration can not be delayed. Direct restoration on the film is a big risk, and the digital processing method is introduced to the video repair technology with its advantages of convenience and repeatability, but comparing with digital image, digital video’s structure is more complex, and its data volume is also huge, so the digital video’s restoration is more complex, in recent years, the restoration of the digital video has become a difficulty and hotspot in digital image processing field.This article mainly aims at researching digital video restoration technology of old film material, the main work and innovation are as follows:(1) There is a deep research and analysis for existing digital image and video restoration technology. Existing digital image inpainting methods are mainly based on the total variation image inpainting algorithm and texture synthesis digital image inpainting algorithm. The former is faster, but it’s just fit for smaller damages in digital image; the latter can restore the bigger damages in digital image, but it has slower speed. The current digital video restoration algorithm is almost evolved from digital image inpainting algorithm.(2) Through deep analysis of a large number of old film materials’damage characteristics, find that most of the damages are small patches and horizontal or vertical scratches, and a few of the damages are large area damages, and the former mostly has strong time continuity, the latter has few time continuity, This paper puts forward an algorithm to distinguish this two types of different damaged objects:by calculating the vertical and horizontal mean widths of damaged area are whether exceed the predetermined threshold to distinguish these two types of different damaged objects.(3) This paper proposes a digital video restoration algorithm based on distinguishing damaged objects, this algorithm restore different damaged objects with different methods:for the characteristic of the first kind of damages, it’s suitable to use total variation image inpainting algorithm to restore; for the characteristic of the second kind of damages, it’s suitable to use the digital video restoration algorithm based texture synthesis to restore, and it’s effect is satisfactory. The proposed digital video restoration algorithm combines these two algorithms, and optimizes and improves these two algorithms, its restoration effect of old film materials is satisfactory.(4) This paper put forward simulation for above digital video restoration algorithm by Matlab7.1, and analyzed and compared deeply the results of the experiment. Experimental results show that the algorithm can work well to restore the damages in old film materials, and get an ideal result.