轮廓跟踪的TN方法
A TN Approach for Contour Following
-
摘要: 在灰度图像二值化时,断边往往会导致伪边界的生成,因此在边界检测时断边处理是非常重要的.提出一种可处理断边的轮廓跟踪方法,在遇到断边沟时,通过自适应步长,能自动跳过断边沟并跟踪出真正需要的轮廓.该方法还具有边界上采样点的疏密度与边界的曲率相适应的优点.Abstract: Boundary detection is of great significance in image processing.Broken edges often result in spurious boundary when an image is binarized.A new contour following approach is proposed to deal with the broken edge gaps.It makes possible to skip the gaps and track out the real contour by means of taking adaptive steps.The boundaries are output in forms of polygons which have the advantage that the density of sample points is closely related to the curvature of boundaries.
下载: