摘要
基于小波变换的奇异值检测能力 ,提出了一种新的平面曲线角点检测算法 .使用高斯函数的一阶导数作为基本小波 ,对曲线的方向角函数进行多尺度小波分析得到候选点集 ,提出了四个筛选规则对候选点集进行逐步筛选得到正确角点集 .仿真试验说明 ,本文的方法不仅能够有效地检测平面曲线角点 ,而且具有较强的抑制噪声能力 .另外 ,此算法简单、易实现 。
Based on the singularity detection ability of the wavelet analysis, a new algorithm to detect the corner points of planar curves is proposed in this paper. Firstly, using the first order derivative of Gaussian function as the base wavelet, the direction angle function of a planar curve was analyzed on multiple scales to obtain the candidates of corners. Then, four filter rules proposed in this paper were applied to the candidates sequentially to get the real corners of the planar curve. Computer simulations prove the high efficiency and strong de noising ability of the new detection method. And because the filter rules are simple and easy to be realized, this new algorithm is also a practical research result.
出处
《小型微型计算机系统》
CSCD
北大核心
2003年第11期1938-1941,共4页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目 ( 60 1740 3 0 )资助
高等学校博士学科专项科研基金 ( 2 0 0 10 6980 0 1)资助
关键词
小波
阈值
角点检测
wavelet
thresholding
corner detection
