Research On Epitaph Digital Tubing Recovery Method

Tombstone,as a carrier of ancient Chinese cultural relics,not only records the life of the characters,but also reflects the political and cultural economy at that time,and plays an important role in the study of history.Tombstones existed during the Wei and Jin Dynasties,and they developed most during the Sui and Tang Dynasties.As time goes by,many tombstones with important historical significance have suffered different degrees of damage.Many of the epitaph digital rubbings obtained today are of low quality,with broken surface characters and a large number of plaques(noise).Therefore,this paper takes the restoration method of the epitaph digital rubbings as the research topic,in order to improve the quality of digital rubbings through its own research.The tombstones of the Sui and Tang Dynasties developed rapidly and were rich in content,and passed dowm in large number.This paper focus on the digital rubbing of epitaph complilation in Sui,Tang and five dynasties.The main work includes the following aspects:(1)Explain the research background of the restoration of the digital epitaph and the significance of carrying out this topic.The current research status of text restoration work is analyzed.(2)Introduction about the relevant recovery methods includes image enhancement,image segmentation and evaluation methods.In terms of image enhancement,it is divided into two aspects: spatial domain enhancement and frequency domain enhancement.Spatial domain introduces linear transformation,piecewise linear transformation and nonlinear transformation;frequency domain enhancement introduces low-pass filter,high-pass filter and homomorphic filter.Image segmentation is mainly explained from three aspects: threshold,region and detection.The evaluation method of image restoration is introduced from two aspects of subjective evaluation and objective evaluation.(3)Analyzed the noise cause and text distribution characteristics of digital rubbing,and introduced the noise probability density function.In the image preprocessing stage,the Retinex method is used to enhance the dark areas of the text to improve the contrast between the text and the background area.(4)A BM3 D algorithm based on block image information entropy is proposed.The relationship between block image information entropy and image noise is analyzed.The BM3 D algotithm is used to denoise noise.Different threshold is used for different classes.Experiments show that this method can protect the strokes of text while removing most of the noise.(5)The quality of the damaged digital rubbings was further improved by using the modified Canny edge detection operator and the CDD model.Firstly,the iterative binarization method is used to convert the digital rubbing into binary image;then the Canny edge detection operator is used to locate the text region of the digital rubbing piece;The text area is etched by mathematical morphology,and the noise image of the digital rubbing is obtained by the difference method.The noise image of the rubbing film(that is,the mask of the CDD model);using the obtained mask to automatically repair the digital rubbing.The proposed method is evaluated from both subjective and objective aspects.Experiments show that the proposed method can fix certain background scratches and remove background noise,and can protect the text structure from being affected.