【摘要】 针对轮廓曲线的多边形近似和特征点提取,提出了多边形逼近误差和局部最小误差逼近特征点的定义和相应的实现算法。该特征点对轮廓曲线进行树状递归划分,并最大限度地减小逼近误差。使得在给定特征点数目情况下,多边形逼近误差为最小。在给定逼近误差的情况下,特征点数目为最少。对于轮廓线的特征提取、优化多边形逼近、压缩表示具有一定的意义。更多还原

【Abstract】 An algorithm was developed to extract outline feature points and to approximate polygon using a polygon approximation error and a local minimum approach error for the feature points.The feature points divide the outline into a tree form with the minimum approach error,using the least number of feature points for a given approach error and minimum approach error.The feature points are used to feature extraction,optimization of polygon approximations,and compression of the outline curve.更多还原