Research On Bitmap Vectorization Technology

With the development of human society, the graphics have been deep into every aspect of people’s lives. From the pictures in books to magazines, from video images to the electronic map on the Internet, from fashion figure to industrial drawing and advertising poster on streets and lanes, it plays an important role in industry and graphics in the industrial, medical, IT filed. Through the combination of computer technology and industrial design, the use of CAD technology has made a fast development. People can input workpiece parameters simply by computer or by scanning of the workpiece then the production can be done by computer-controlled machine. In that way the work can be greatly improved. At the same time, early design drawings need to be preserved due to the impact of the preservation of the environment. However, saving a huge number of design drawings and data is bound to spend a lot of manpower and resources. Through the entered automatically with the intelligent recognition technology, can effectively solve the above problems. It is easy to facilitate the data, accelerate the upgrading and development of the product.This article first introduces the image pre-processing technology, exemplified the characteristics of the graphics. Introduces figure vectorization process in the first step of image preprocessing. This paper introduces detailed about digitized, geometric transformation, normalization, smoothing, restoration and enhancement steps. And separated threshold selection, the process of image restoration, image enhancement method for image smoothing method to do a detail analysis and compared the advantages of different methods.Secondly, this paper describes the vector generation and general image vectorization step details repetition of the boundary, the distance to the exchange method, appropriate skeleton, based on the Hough transform, contour-based approach, based on sparse pixel method commonly used image algorithm vectorization traditional. Introduce the work steps of sparse pixels method, by comparing the advantages and disadvantages of various algorithms, we derived algorithm suitable herein-based Hough transform the improved method, i.e. before the classic Hough transform, make the first image pre-processing, such as polygon extract, or a cubic Bezier fitting to obtain the coordinate data.Finally, we give a detailed instructions based on the improved Hough transform data, then a brief statement has been made focus on the development environment and design principles. Using matlab, This paper design object-oriented operating software. Vector processing carried out on the sphere and pentagram using polygon extraction has got their vector pictures; vector processing conducted on human skull and the sphere using Bezier fitting has got both of their vector pictures. Both methods are significantly improved image aliasing. At the same time, respectively made the application by extracting polygon method and Bezier fitting, and do a comparison test.The result of the proved polygon extraction method have better effect by straight lines, and the Bezier contemplated lawful curve effect is better.