期刊文献+

数字曲线上特征点的检测 被引量:7

ON FEATURE POINT DETECTION OF DIGITAL CURVES
下载PDF
导出
摘要 本文提出了一种检测数字曲线上特征点的算法.它首先总结了目前已有的多边形近似和显著点检测算法,然后指出正确的特征点检测算法应结合估计和形状两方面的准则,最后提出了一种基于可靠线段的特征点检测方法.这种方法先从估计的观点找出曲线上的可靠线段,然后从形状理论的观点来组合这些线段以及剩余边缘点来获得曲线的特征点.如何寻找曲线上的可靠线段,本文提出了基于特征图分析的方法和卡尔曼滤波的方法.实验结果表明这种方法是有效的. An algorithm for detecting feature points on a digital curve is proposed.The paper first summarizes some current approaches to polygonal approximation and dominant point detection, then points out that the accurate algorithm for feature point detection should combine the principles in the approximation and shape theory, and finally develops a procedure based on reliable lines for detection of feature points. This procedure first finds reliable lines from an approximation theory point of view, and then from a shape theory point of view combines these lines and residue points to acquire feature points on a digital curve. The paper develops two approaches to find reliable lines on a curve. One is based on feature graph analysis,another is based on the Kalman filter. The experimental results show the algorithm is very efficient.
出处 《计算机学报》 EI CSCD 北大核心 1998年第6期520-526,共7页 Chinese Journal of Computers
基金 国防预研基金
关键词 特征点检测 数字曲线 卡尔曼滤波 图像分析 Feature point detection, polygonal approximation, reliable line, feature graph analysis, Kalman filter
  • 引文网络
  • 相关文献

参考文献3

  • 1文贡坚,硕士学位论文,1997年
  • 2Zhu P,IEEE Trans Pattern Anal Mach Intell,1995年,17卷,8期,737页
  • 3张有为,维纳与卡尔曼滤波理论导论,1980年

同被引文献81

引证文献7

二级引证文献33

;
使用帮助 返回顶部