基于分段与识别技术的平面轮廓的精确重构

Accurate Reconstruction for Planar Contour Based on Segmentation and Recognition Techniques

解决了由直线和圆弧构成的平面轮廓精确重构问题,满足轮廓各段之间位置连续以及相切条件。在正确识别角点、切点等特征点的基础上,提出了约束最小二乘算法,可精确求解带边界约束圆弧的重构问题。将圆弧边界约束条件分为5种类型:单点约束、双点约束、单切矢约束、单点单切矢约束和双切矢约束。针对每种类型给出了具体的计算公式。详细描述了轮廓精确重构的算法流程,可实现在全局误差控制下的精确重构,最后给出了运行实例,验证了算法的有效性。

Accurate reconstruction of planar contour mixed with straight lines and arcs is achieved with position or tangent continuity. Based on segmentation and recognition technology for planar contour, constraint least square （CLS） algorithm is proposed to reconstruct arcs with boundary constraints. CLS is realized under the situation of five types of boundary constraints： single comer point, double comer points, single tangent point, single comer point, single tangent point and double tangent points. To avoid accumulating error caused by reconstruction with boundary constraints, an optimized algorithm is discussed to control the global error. Some examples are presented at last.