Mural Pigment Identification Method Based On Multispectral Imaging Technology

Mural painting is an important cultural relic resource in china.However,due to the destruction of the ancient times and various environmental factors,many murals have different degrees of blurred images,pigment off and other issues,which affect the preservation of the murals.In order to solve this problem,there are a lot of cultural relic workers dedicated to the study of the mural restoration.The traditional method of pigment recognition is inefficient and destructive,and the multi-spectral imaging technology has the advantages of efficient,accurate,and nondestructive.According to the characteristics of spectral reflectance data,in this thesis the spectral matching algorithm based on improved edit distance is proposed,and is used to recognize a mural pigments in a museum of Xianyang,on the basis of the ancient wall painting pigments intelligent analysis system is designed.Specific work as follows:Firstly,a spectral matching algorithm based on improved edit distance is proposed.The traditional edit distance algorithm for spectral matching,are sensitive to the amplitude,the matching accuracy is low.Therefore,the use of adaptive threshold instead of zero threshold conditions as the decision condition of edit distance algorithm can solve this problem.At first calculation sample and measured reflectance difference of pigment by point,forming a difference vector,and then calculate the threshold of difference vector using category variance,and count the number of the points in difference vector,whose value is the greater than threshold,the statistical result is the edit distance of two reflectivity curves,and finally determine the recognition result according to the ratio of the edit distance and data points.The experimental results show that the matching precision of improved algorithm is better than the traditional edit distance algorithm,can solve the problem of low accuracy of traditional spectral matching algorithm to identify the metamerism pigments,the method can accurately identify the mural pigment.The intelligent analysis system of ancient wall painting pigments is designed.The system integrates the color classification module,storage and output module and analysis of pigment mural pigment sample database module,realize the automatic classification of the ancient wall paintings of different color regions,and canrecognizethepigmentsin each colorregion.The system can protect the digital filing of the murals,provide the information of the digital preservation of the murals,and provide guidance for the restoration of the color of the murals.