| Based on the theory of digital image processing, this thesis makes further researches into image restoration according to the technique of wavelet transformation on the background of monitoring system for Dayaoshan Tunnel, especially introduces the application of basic wavelet theory in image denoising and restoration of motion blurred images.First, the purpose and meaning of item research are presented, and the project background offers practical foundation to the building of image models and theoretical researches. Then, in the facet of image denoising, the thesis brings up a new wavelet algorithm based on wavelet shrinkage, which is called improved wavelet denoising algorithm via empirical wiener filtering. Generally, wavelet denoising algorithms choose the threshold just by experience which has a great influence on the result of image denoising. However, the thesis combines the wavelet technique and design of the wiener filter together considering the effect of noises in the tunnel, and estimates the parameters of image and noise effectively that will be used to design the wiener filter. In that case, it will be able to improve the image performance and experiment result. Moreover, it also will monitor the important places'safety more effectively.Considering the appearance of motion blurred image in the process of collecting images of tunnel, the thesis builds the degeneration model of motion blurred image and gives a new way to identify the motion blur direction form the blurred image by direction derivation method in different signal noise ratio. The new way not only improves the identified precision of image in low signal noise ratio, but also heightens performance against noise. In addition, several representative methods about image restoration are analyzed systematically. Finally the thesis makes a summary and research prospect. |
PDF Full Text Request