高分辨率图像特征提取和VG重构技术研究  

RESEARCH ON FEATURE EXTRACTION OF HIGH RESOLUTION IMAGES AND VG RECONSTRUCTION

作　　者：童庆[1] 张敬谊[1] 陈诚[1] 张立鹏[2] 顾春华[2] 何高奇[2] 

机构地区：[1]万达信息股份有限公司研究发展中心,上海201112 [2]华东理工大学计算机科学与工程系,上海200237

出　　处：《计算机应用与软件》2013年第7期116-118,184,共4页Computer Applications and Software

基　　金：上海市自然科学基金项目(10ZR1408200);地理信息科学教育部重点室验室开放研究基金项目(KLGIS2011A06)

摘　　要：高分辨光栅图像的存储和实时传输是当今图形图像研究的一个热点。通过将高分辨率数字图像矢量化,显著降低图像存储所需的空间和图像传输的带宽要求,提高网络环境下图像信息传输的完整性和绘制的实时性。首先对采集得到的图像数据进行预处理,然后基于改进的LoG(Laplace of Gauss)算子提取其中的边缘信息。接着根据边缘信息,利用Delaunay对图像进行三角化,并提取三角区域的颜色信息。最后,将三角化后的数据和相应的颜色信息保存为XML格式,形成SVG文件来实现网络环境中的传输和重绘。实验结果表明,利用其使用的方法可以实现高分辨图像在不同客户端下的实时显示。The storage and real-time transmission of high resolution raster images are the hot subject in images and graphics researching field. By the vectorisation of high resolution digital image, the storage space of images and the bandwidth requirements for image transmission are significantly reduced, this improves the integrity of image information transmission and the real-time performance of image rendering in network environment. In this paper, we first pre-process the collected image data, and then extract the edge information using the modified LOG operator. After that, according to the edge information, we use Delaunay to implement the triangulation on the images, and extract the colour information in triangular areas. Finally, all these triangulated data and corresponding colour information are written in the XML format to form an SVG file for transmission and re-rendering in network environment. Experiment results prove that the use of the method proposed in the paper can well realise the real-time display of high resolution raster images at different clients.

关 键 词：特征提取 高斯拉普拉斯算子 矢量化 三角化 可缩放矢量图形 绘制 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]