Research On Restoration Technology Of Damaged Tibetan Documents Based On Image Processing

In real life,due to the limitations of the inherent hardware equipment and natural environment of the imaging system,people often can not get clear images.It costs too much to get high-resolution image by improving hardware equipment,and it is difficult to overcome the technical bottleneck for a specific imaging system in a short time.With the development of computer science and the maturity of signal processing technology,the resolution can be improved by digital image processing technology.As an important research direction of digital image processing,image restoration technology has been paid more attention in recent years.Based on the research results of scholars at home and abroad,this paper briefly introduces several classic image restoration algorithms,analyzes the shortcomings of these algorithms combined with the research object of this paper,and puts forward corresponding improved algorithms for the existing problems.This paper focuses on the practical problems in the damaged Tibetan ancient books and documents,and devotes itself to the restoration with digital image processing technology.The main research contents include the following aspects:First of all,the collection of Tibetan ancient books and documents for professional image collection and preprocessing.The professional CCD camera and image acquisition card are used to collect samples,which ensures the accuracy of image acquisition.The preprocessing work mainly focuses on three aspects:geometric correction,edge detection and image filling,which lays the foundation for the follow-up algorithm research.Secondly,an improved algorithm based on nonlocal mean filtering is proposed to recover theL₀ norm text image.The image gradientL₀ norm is used as the regular term to constrain the filtering restoration process and estimate the fuzzy kernel,so as to solve the problem of degradation in Tibetan ancient books.In the algorithm framework of non-local mean filtering,the calculation method of weight function is improved to reduce the calculation amount.The simulation results show that the improved non local mean filter restoration algorithm has a good performance of comprehensive quality evaluation in text image restoration.Finally,an improved text image super-resolution restoration algorithm is proposed.This algorithm is based on the maximum posterior probability estimation framework of Markov random field model,fully combined with the statistical characteristics of text image bimodal distribution,which not only guarantees the restoration performance of the algorithm,but also prevents the distortion caused by the image over filtering.Through contrast experiments and image quality evaluation,the results show that the improved algorithm in this chapter has a good recovery ability,no matter in subjective visual perception and PSNR,SSIM objective quality evaluation.