一种基于分裂合并的多边形逼近算法

Polygon Approximation Algorithm Based on the Split and Merge

断层医学图象目标组织经图象分割、轮廓跟踪后得到的轮廓像素点数据量较大,不宜直接用于几何建模。多边形逼近是提取曲线特征点和简化数据、加快图形运算的一个重要方法,提出一种基于分裂合并的多边形逼近算法,将轮廓像素点集合分段进行线段逼近,逐次递增进行共线检查,反复执行分裂、合并操作,直到所有逼近误差在指定范围内,逼近多边形不再改变为止。该算法能够在保持原始轮廓形状特征的前提下,有效减少数据量,提高了计算效率。

The outline pixels of sectional medical image goal organization are not appropriate to be directly used in the geometric modeling,for it is with large amount of data after image segmentation,contour following.Polygon approximation is an important method in feature points extration and data reducing.In general,it presents a new polygonal approximation approach based on the fission and fusion,it will segment the outline pixel for line approximation and check collinear successive increase,fission and fusion,until all the approximate error in the designated area and approaching polygonal is no longer changed.The method can reduce the quantity of data effectively without changing the original profile,and greatly increase the computation efficiency.