邹静, 陈永刚, 龚金琪, 董万虎, 孙燕飞, 王志林. 一种利用多维目标分割比的矢量图形匹配算法[J]. 武汉大学学报 ( 信息科学版), 2020, 45(10): 1626-1632. DOI: 10.13203/j.whugis20190009
引用本文: 邹静, 陈永刚, 龚金琪, 董万虎, 孙燕飞, 王志林. 一种利用多维目标分割比的矢量图形匹配算法[J]. 武汉大学学报 ( 信息科学版), 2020, 45(10): 1626-1632. DOI: 10.13203/j.whugis20190009
ZOU Jing, CHEN Yonggang, GONG Jinqi, DONG Wanhu, SUN Yanfei, WANG Zhilin. An Efficient Matching Algorithm Based on Vector Graphics Using Multi-dimensional Object Segmentation Ratio[J]. Geomatics and Information Science of Wuhan University, 2020, 45(10): 1626-1632. DOI: 10.13203/j.whugis20190009
Citation: ZOU Jing, CHEN Yonggang, GONG Jinqi, DONG Wanhu, SUN Yanfei, WANG Zhilin. An Efficient Matching Algorithm Based on Vector Graphics Using Multi-dimensional Object Segmentation Ratio[J]. Geomatics and Information Science of Wuhan University, 2020, 45(10): 1626-1632. DOI: 10.13203/j.whugis20190009

一种利用多维目标分割比的矢量图形匹配算法

An Efficient Matching Algorithm Based on Vector Graphics Using Multi-dimensional Object Segmentation Ratio

  • 摘要: 图形的旋转、缩放、平移(rotate, scaling, translation, RST)以及边界微形变是影响矢量图形匹配结果的主要因素, 也是评价形状描述算法优良的标准。针对以上因素, 提出了一种多维目标分割比的矢量图形匹配算法。该算法通过对矢量图形构建特征线的方式实现形状特征的提取, 准确描述面目标要素的形状特征, 度量要素间的形状相似性。以2 000个真实地理实体的矢量图形作为匹配标准库, 随机选取半数图形作为待匹配图形, 对其采取不同程度的RST变换和边界简化, 然后与标准库做匹配试验, 并与其他图形匹配算法进行对比。试验结果表明, 所提出的算法具有更高的匹配准确率, 具有RST不变性和形变鲁棒性, 可以精准识别矢量图形的形状。

     

    Abstract: Rotation, scaling, translation (RST) and boundary micro-deformation are the main factors that affect the matching results of vector graphics, and are also good criteria for evaluating shape description algorithms. In this paper, a vector graphics matching algorithm using multi-dimensional object segmentation ratio is proposed for the influencing factors of shape matching. This algorithm extracts shape characteristic by constructing feature lines for vector graphics. It can accurately describe the shape characteristic of the polygon features to measure the shape similarity between the graphics. The vector graphics of 2 000 real geographical entities are used as the standard library for matching experiment, and half of the graphics are randomly selected as matching graphics. The matching graphics are taken to RST transformation and boundary simplification operations with varying degrees. The transformed graphics are matched with the standard library. The matching results are compared with other vector graphics matching algorithms to test the shape retrieval effect of this algorithm. This experiment demonstrate that the proposed algorithm has higher matching accuracy, RST invariance and deformation robustness. Therefore, it can accurately identify the shape of vector graphics.

     

/

返回文章
返回