摘要
轮廓曲线包含了图像中大量的关键信息,主要用于探讨基于轮廓曲线的图像匹配问题.在假定图像间具有相似关系的前提下,借助于一种完备正交函数系V系统,提出一种新的曲线匹配算法.首先由V描述子评估曲线间的相似度,以此确定匹配曲线段,然后求得曲线间的变换参数,最后进行了图像拼接测试.该算法着眼于轮廓线的整体特征且不需要提取曲线的局部特征,避免了特征选取这一难点.大量的模拟及真实图像实验结果表明,从图像中提取的2条曲线间即使不严格满足相似变换关系,依然可取得比较满意的结果,从而验证了文中算法的可行性.
Considering that contour contains a lot of critical information in the image,a new curve matching algorithm is proposed based on the V-system which is a class complete orthogonal function system under the assumption that the two images are similar.Firstly,the distance is calculated between curves using the descriptor of V-system to find the matching curves,and then the matching parameters are evaluated,Finally,the two images are stitched together.Different from the existing approaches,this paper considers contour features as a whole feature curve,thus the difficulty of feature selection is avoided.Through a large number of simulations and real images,in spite of the fact that the images do not satisfy the similar transformation strictly,the result is still satisfactory which verifies the feasibility of the algorithm.
出处
《计算机辅助设计与图形学学报》
EI
CSCD
北大核心
2012年第4期471-477,共7页
Journal of Computer-Aided Design & Computer Graphics
基金
国家自然科学基金(60973005)
北京市教育委员会科技计划面上项目(KM201010009008)
关键词
完备正交系
V系统
曲线匹配
相似变换
complete orthogonal system
V-system
curve matching
similar transformation
