Application Of Digital Image Restoration Technology In Recovery Of Rock Paintings In Donkey Gully

Rock paintings are ancient paintings of humans on the stone and are important materials for the study of Chinese history.However,with the passage of time and the ages,various natural factors and human factors have greatly affected rock art.Many rock paintings have been damaged to vary ing degrees.Some rock paintings appeared cracks,and some rock surfaces fell off,causing extensive damage to rock paintings.Therefore,it is imperative to repair rock art.Human-made manual repair is easily affected by personal thinking,and it is not conducive to preservation.In recent years,digital image technology has developed rapidly.Applying digital image restoration technology to the protection of rock art can yield very impressive results.This paper proposes a set of repair methods for damaged rock paintings in the donkey trench area.Through experiments,the proposed method is feasible.The main work of this paper is as follows:(1)Preprocess the collected donkey drain rock paintings.Because the rock-painted images directly used for denoising are directly used for image restoration,some features of the rock itself,i.e.,rock picture images are not clear,and the unevenness of the rock surface is unequal,so that the repair result is not ideal.In order to solve the above problems,this paper proposes a method of preprocessing rock picture for image denoising and image grayscale transformation.In terms of denoising,this article mainly introduces the basic principles of several commonly used denoising methods,namely the mean filtering and the median filtering,and obtains the application of the two denoising methods through experiments.In the aspect of image gradation conversion,the linear gradation transformation method is first introduced.,and then introduce the relative non-linear gray transformation method,as well as the basic principle of histogram homogenization.The experimental results show that the denoising and gradation transformation processing of the collected image in advance help to improve the clarity of the image and protect the useful information of the image,which is helpful for the subsequent repair of the results.(2)Extract broken areas of rock paintings.Before the repair of rock paintings,it is necessary to extract the damaged areas.This paper proposes an automatic crack detection method based on the knowledge of high-hat computing and expansion calculations.The top cap is replaced first,then the threshold segmentation is performed,then the morphological filtering method is used to process the details,and finally the denoising operation is performed.In this paper,extraction of cracks from structural images can achieve full-automatic crack extraction,and the method of combining artificial marking with automatic extraction of rock-painted images with complex textures can be used to extract damaged areas of rock paintings.The method in this article can improve the defects of personal thoughts in the hand-tagged damaged area.(3)An improved image restoration algorithm based on TV model is proposed for the repair of small crack rock image.Usually,image restoration algorithms based on TV model are used to repair small crack images,and certain repair effects are obtained.At the same time,there are also some deficiencies,such as sensitivity to parameters and poor robustness.In addition,the traditional TV algorithm to repair slightly damaged range of images prone to blurring and staircase effect.In order to overcome the above defects,this paper proposes an improved rock painting repair algorithm based on the nonlinear diffusion theory.This algorithm can effectively solve the problem of the failure of the traditional TV repair algorithm in small areas.Comparing the repair results of the relevant algorithms through experiments shows that the proposed algorithm achieves a better repair effect compared to the traditional TV model.(4)An image restoration method based on image decomposition was proposed for texture rock paintings with extensive damage.Usually,the classical image restoration is the Crimeinisi algorithm,which has a good effect on the general large-scale damage of the texture image repair,and it is lacking in meeting the visual connectivity requirements.In view of the above problems,this paper proposes an image restoration method based on image decomposition.Its basic idea is to first decompose the image into structural and texture parts.The structural part adopts the improved image algorithm based on TV model proposed in this paper,and the texture part is repaired by the improved Criminisi image repair algorithm.Finally,the restored structure image and the restored texture image are reassembled into a new image to obtain complete rock painting repair.image.Experimental results show that the proposed image restoration method can improve the sharpness of texture image restoration,protect the texture information,and satisfy the visual connectivity.