Abstract:

A new method named quadric surface direct fitting based on normal vectors of measured data is proposed to overcome the current quadric surface fitting method’s shortcomings. In the current method, the calculations of the initial geometric parameters are imprecise, and the using of it in surface fitting can cause poor efficiency and stability. Then, the performance contrasts of this method with existing methods are presented　based on various test models.

Key words: computational , geometry,  , quadric , surface,  , reverse , engineering,  , surface , fitting,  , feature , extracting