Research And Implementation Of Multi-scale Vectorization Technology For Raster Images Based On Contour Lines

Raster image vectorization is a hot issue in the field of digital image processing,which is widely used in terms of digital image processing,editing and animation production,etc.With the development of network communication technology and the popularity of handheld terminals,the use of mobile devices such as smart phones and iPad is increasing,so more and more people put forward higher requirements on the storage and editable of digital images.In order to meet the practical needs of multimedia applications in recent years,scholars in the field of image processing began research on color image vectorization.Raster image vectorization needs not only the boundary information of raster images,but also requires internal geometry and color information of the image.In this regard,based on the existing raster image vectorization algorithm research,we propose a novel algorithm for image vectorization by using contour-based multi-scale representation in this paper,the main contents include raster image preprocessing,raster image vectoriza-tion and reconstruction of raster images.The main work of this paper include the following aspects:1.Raster image preprocessing.Firstly,by converting the RGB values of the raster image to a corresponding gray value obtain grayscale raster image;Then use the smoothing filter operator for grayscale smoothing,filtering image noise information and soften the image color mutations.Experimental results show that the algorithm can quickly perform a variety of raster image preprocessing,which can be treated effectively applied to raster image vectorization process.2.Contour-based raster image vectorization.Upon completion of image preprocessing,the next step is to raster image vectorization.The first step in this phase is to extracts contour points,which is taken as vectorization samples of the original image,from the grayscale image.And extract the corresponding color information and geometric information of the contour points,which can be used as the original information of raster image vectorization.Further,the algorithm constructs a set of contours defined by the samples.By reverse-calculated control points of B-spline curves,the mathematical representation of vectorization image is generated by data approximation method.Experiments show that the algorithm is stable,robust,and it can be used in the vectorized process for various digital images.3.Raster image reconstruction based on the Laplace equation.In order to meet the requirement for drawing vector graphics,this paper also provides a method of raster image reconstruction,which is based on the data of vector image.Firstly,by scaling the control points of b-spline curve proportional according to the requirements of users,it's able to control the vector image scaling;And use based on the cubic B spline curve midpoint algorithm of raster display rasterization processing for the color of the contour curve;Then by solving Laplace equation was carried out on the color curve after the rasterize diffusion iterations,obtained by vector said reconstruction of raster image.Experiments show that the algorithm can quickly realize the reconstruction of the raster image,the reconstruction of raster image distortion.In the above algorithm based on theoretical research,this paper implements the algorithm for image vectorization by using contour-based multi-scale representation.Through testing the part of the raster image cases,results of the programs show that the algorithm can not only generate the vector image of the multi-scale,also provides a use the method of parameter control vector image data quantity size.