Parameters Estimation Of Blur Image And Study On Image Restoration

Nowadays, with the general acknowledgement of computers, the digital image processing (DIP) technology is widely being applied in various fields. Of course, the DIP technology is gradually being consummated and systematization. Digital image restoration, as an important research content of DIP technology, has produced lots of research results, which have widely been applied in scientific research, industrial and agricultural production, military technology, government department, health care, etc. With the updating of optics imaging facility, the quality of graph is enhancing. but the clear image come into being just at the narrow range of focus plane. The scenery contains abundant space information in most graphing. So the blur in part can not avoid. This phenomenon is disbennifit to some work, such as cytochemistry and microphysic. Thus, the research about DIP technology has its considerable practical meaning and practicality value.For the blur system without any apriority information to utilize, the traditional methods, as image estimation method, test estimation method and model estimation method, can't obtain the parameters of the blur system. The research in-depth was made in image definition estimation, point spread function's parameter judgment and blurred image restoration techniques. To restore this type blurring image, the author finishes the research on the estimation of image quality and the estimation of blur system. This lay a foundation for image restoration. The main tasks in this paper:①A novel method about image definition estimation automatically appears. The method was based on the gray gratitude of image. It succeeds the good qualities of traditional methods, and the estimation function was added a threshold on. Compare to the methods ago, it has some advantages of single peak, big slope at both sides of the peak. It can finish the searches of legible image quickly. This method can be used in optics imaging system.②In most methods of image restoration, the more priori knowledge we know, the more perfect restoration results we can obtain. Thus, it is an important work to estimate the parameters of blurring system. But according to the experience of the relative work before, there is not a good way to estimate the parameters of blurring system yet. In the paper, the author raises a new method of parameter estimation, which is based on the estimation of restored image quality. What's more, the emulational tests proved that this method could estimate the parameters of the point spread function accurately, especially in the conditions of good Signal-to-Noise.③The author makes a research in-depth at the theory of Wiener filter and Restricted auto-adaptive restoration method(Lagendijk,1988), and has accumulated plenty of experience and summarize that most methods make a compromise between deblurring the edge and the flat. The Restricted auto-adaptive restoration method makes restrictions on both of the edge and the flat, and obtains perfect results relatively, while this method has some disadvantages. In this paper, using the idea of the restricted theory and combining it with the Wiener filter, the author brings out a new method based on the restricted auto-adaptive restoration method.At last are a summary and a blueprint.