Research On The Edge Extraction Of Huizhou Woodcarving Images

Huizhou woodcarving is known for its superb carving technology and deep cultural deposits.Huizhou woodcarving is called the best products of human culture and art because it is beautiful in shape,exquisite in structure and complex in texture.It is very necessary to use digital technology to protect and promote the intangible culture.The edges extracted from woodcarving images can be used for vectorization and can be applied to architecture,furniture,products and other related fields,which is conducive to the inheritance and protection of Huizhou carving heritage,and the secondary development of woodcarving images.However,there are many complex texture details in Huizhou woodcarving images and the edge gray gradation is blurred.The edge images extracted from the traditional edge detection methods contain a lot of redundant details and false edges,it is not conducive to pattern recognition.In this thesis,the traditional Canny algorithm is improved,and a prototype system is developed for the edge extraction of Huizhou woodcarving images.The main work of this thesis is as follows:(1)The Gaussian filter used in traditional Canny algorithm is difficult to suppress the noise and maintain the edge of the image at the same time.To solve the problem,bilateral filter is used to replace Gaussian filter to smoothe the wood grain while preserving the edges and denoising.The finite first order difference is used to calculate the gradient of the gray value of the image,and the gray information of the neighborhood pixels is not fully considered.To solve the problem,a Sobel template is used to replace the original template to contain more gray information of the neighborhood pixels.The experimental results show that the improved Canny algorithm can effectively smooth the wood grain and noise,the extracted contour edges are easy to identify.(2)In order to solve the problem that the gradient operators are insensitive to the blurred edges of Huizhou woodcarving images,an improved edge detection method based on fuzzy reasoning is adopted in this thesis.Firstly,twelve gray difference values of each pixel in its eight neighborhood are calculated as inputs.Then,two membership functions are used to convert the exact values into fuzzy values,and the fuzzy operators are used to get the edge membership of the center point pixel by 20 fuzzy rules.Finally,if the edge membership is greater than the threshold,the pixel point is marked as the edge point.The experimental results show that the improved method can effectively enhance fuzzy edges and weak edges of woodcarving images.(3)A prototype system is developed for the edge extraction of woodcarving images.The system can import the original image,display the gray image and the gray histogram to reflect the gray information of the image.Roberts,Sobel and some other classical edge detection operators can be chosen to detect the edges,while providing improved Canny algorithm and improved edge detection algorithm based on fuzzy reasoning.In this system,some important parameters can be manuallyadjusted and the results can be immediately observed,the extracted edges can be thickened or refined.Finally,the edge images can be saved.