摘要
提出基于距离函数的运动生成曲面的识别和重构的新方法。通过分析与距离函数无关的微分运动所组成的线性空间的特征来识别曲面的类型,并由此计算出有关参数的初始值,将B样条曲线的控制点作为曲面形状的控制因子,基于有向距离函数的微分性质,应用Levenberg-Marquardt算法精细调节待重构曲面的空间位姿和形状。该方法能够自动识别简单曲面和运动生成曲面的类型并重构出曲面的特征,回转曲面和螺旋面的识别重构的仿真试验结果表明该方法具有很高的精度。
A new method of identification and reconstruction of motion-generated curves based on distance function is proposed. The type of curve is identified through analyzing the characteristic of linear space formed by differential motions unrelated with distance function, and thereby the initial values of related parameters are calculated. The control point orB spline curve is taken as the control factor of surface shape. Based on the differential properties, Levenberg-Marquardt algorithm is employed for fine regulating of spatial pose and shape of the curve to be reconstructed. The method can automatically identify the types of simple curve and motion-generated curve and reconstruct the characteristic of curves. The results of simulation test of identification and reconstruction of rotary curve and helical surface show that the method has very high precision.
出处
《机械工程学报》
EI
CAS
CSCD
北大核心
2009年第10期151-158,共8页
Journal of Mechanical Engineering
基金
国家重点基础研究发展计划(973计划
2005CB724103)
国家自然科学基金(50775147)
上海市基础研究重点(07JC14028)资助项目
关键词
点云数据
曲面识别
曲面重构
距离函数
Point cloud data Surface identification Surface reconstruction Distance function
