一种高效的特定边界检测算法

An Efficient Interested Boundary Tracking Algorithm

为了快速准确的检测特定边界,采用边界点跟踪算法检测特定轮廓边界点.通过分析不同算法的特点,提出了基于已知点夹角的边界点预估计算法和基于圆弧邻域扩展的边界点搜索算法,以避免跟踪进入搜索陷阱,提高边界跟踪对边界变化的鲁棒性.实验结果表明,边界点预估计算法可以有效的避免搜索陷阱的影响,正确地跟踪待检测边界;对不同曲率的边界,检测结果合理,边界点分布均匀.

In order to detect the interested boundary fast and correctly, an boundary point tracking algorithm was proposed to detect the specifical edge of the image. At the basis of analyzing different boundary tracking algorithms, boundary point estimate algorithm based on angle and boundary point searching algorithm based on circumferential neighborhood extending were advanced to avoid tracking trap and to improve the robust of the boundary tracking algorithm. The experiment showed that the proposed algorithm could get out of tracking trap on the whole detectable processing, moreover, for boundary with different curvature the algorithm could get correct boundary points which were distributed equably.