The Research On Fitting Of Standard Spacial Quadratic Surface

As an important approach for improvement of productive efficiency and shape design, more and more attention is paid for reverse engineering, fitting of quadratic surface is one of its components. Nowadays, general fitting methods are based on the character equations of quadratic surface to implement. For it is so complicated that a lot of fitting software exist a lot of restrict conditions. To settle these problems, this paper carries on the thorough research of the fitting algorithm.In this paper, data acquisition and processing method is introduced, surface fitting and error estimation methods based on special scattered data is put forward. Jacobi transformation is used to find correlations among coefficients. Then a method to transform general quadratic surface to formal quadratic surface equation is introduced. The location, direction and size of quadratic surface are obtained.At the same time, for the equidistant surface in practical measurement, special search method is applied for automatic segmentation, so the normal vector of the data at arbitrary position of the space is solved, then radius compensation of probe is implement. So the return is realized from equidistant surface to real measurement surface.Different error estimation methods are put forward. Error calculations are also completed very well. It provides very reliable theoretical basis for production and practice.Massive simulation experiments have proved the correctness and possibility of the fitting algorithm of quadratic surface which is based on general equation.