摘要
提出了一种改进形状上下文特征匹配的线要素Morphing方法。该方法通过对小比例尺线要素进行点加密,使得点对之间的相似性评估不受点集数目差异的影响,仅与线上点的分布形态有关,并对原始形状上下文外围空间区域划分进行加密,使其在描述点的形状上下文时能更充分顾及外围点的影响;然后通过Kuhn-Munkres算法使得最终匹配结果在权重之和最大的同时达到全局最优;最后对匹配结果进一步进行调整,通过线性内插实现Morphing变换。试验表明,该方法与原有利用单一形状上下文进行特征匹配的方法相比,能有效避免Morphing变换过程中线要素自相交情况的出现,且在Morphing变换过程中能较好的保持线要素的形态特征。
A linear feature Morphing method is proposed to improve shape context matching.In this method,the point encryption is applied to small-scale line elements in order that the similarity evaluation between point pairs is not affected by the difference in the number of point sets,but only related to the distribution of points on the line.And the surrounding space division of the original shape context is encrypted,so the influence of peripheral points can be fully considered when describing the shape context of points.Then the Kuhn-Munkres algorithm is used to make the matching result reach the global optimum while the sum of weights is the largest.Finally,the matching results are further adjusted,and Morphing transformation is realized by linear interpolation.Experimental results show that the proposed method can effectively avoid the self-intersection of line elements and can better maintain the morphological characteristics of line elements in Morphing transformation process.
作者
李彩霞
翟仁健
杜佳威
万瑞康
LI Caixia;ZHAI Renjian;DU Jiawei;WAN Ruikang(Information Engineering University,Zhengzhou 450001,China)
出处
《测绘科学技术学报》
北大核心
2019年第6期654-658,共5页
Journal of Geomatics Science and Technology
基金
国家自然科学基金项目(41701463)。
作者简介
李彩霞(1994-),女,河南平顶山人,硕士生,主要从事制图综合与空间数据多尺度表达研究。E-mail:lcx58558@163.com。
