The Research On Crack Recognition Of Ancient Architecture Color Painting Based On Curve Analysis

Color painting, as a kind of decorative art of traditional Chinese ancient architecture, reflects the rich cultural heritage of Chinese culture. It benefits the inheritance and the promotion of the culture to analyze the artistic feature of the painting and to discuss the digitized protection. Due to the aging concrete, diseases of varying degrees often occur to painting. Detection and analysis of those diseases would provide practical basis and guidance for the protection and repair. This paper focuses on the analysis and crack detection of tangent circle pattern painting of Summer Palace building.Tangent pattern painting is one of major category of Qing Dynasty official painting, and has a great varieties and widely application. In this paper we analyze its linear features and the characteristics of their distribution and connection, and then sums those up based on digital images, thus provides practical basis for the crack detection. The major contents of this article are as follows:Firstly, for tangent pattern painting, we analyze its linear features based on lots of painting images and related literature. Then, we give the mathematical expression of those features. Finally we summarize and establish the distribution and connection model of linear relationSecondly, for the painting of simple background, use vector method to detect edges, then carry on broken points connection and wipe out small edges to better accommodate the requirement of crack detection. Next, iterate through the edge information, decompose the edges into single lines by continuity analysis. And then we analyze the characteristics of line through curve fitting. Finally, we compare with the results of linear shape analysis, by analyzing every line segment features and relationships, realize the crack detection.Thirdly, for painting of complex background, simple edge detection can not fully extract the edge information. In this article, background analysis is added in image preprocessing to eliminates the background noise and to improve image contrast in order to fit the color gradients with the requirement of edge detection, through this way, crack detection and recognition of painting images with complex background is achieved.According to the experimental results, the method in this article can accurately get the information of linear features and the characteristics of their distribution and connection. The crack recognition based on curve analysis can better handle with the edges of the image. The results show that the method in this article is effective.