一种新的空间曲线匹配算法

A Novel Algorithm on Space Curve Matching

提出一种新的空间曲线匹配算法。该算法对离散的曲线进行光顺、拟合和采样,使用空间曲线微分几何性质计算各采样点的曲率和弗朗内特标架。通过曲率计算匹配点对列表,通过对齐匹配点对的弗朗内特标架进行曲线匹配,得到一个匹配矩阵集。从匹配矩阵集中选出一个最优匹配的匹配矩阵,使得空间曲线的对应点匹配窗口内的距离平方和最小。实验结果证明,算法效率高、鲁棒性好。

A novel space curve matching algorithm was presented. After smoothing, fitting and sampling the discrete space curves, the algorithm calculated Frenet frame and curvature for each sampiing point. The curvature was used to pair the sampling points. Then by registrating the Frenet frames for each pair, a set of match matrix was obtained. We selected the optimal match matrix from it, which minimized the corresponding point distance square sum within the match window. Experimental results demonstrate the algorithm is efficient and robust.