Study on the technology of quadric surface extracting base on least square method

This paper studies quadric surface extracting direct focus on scattered points clouds data. Firstly, the points clouds data of single style are distinguished from curved surface. Then, using the geometric parameter equation of the different style quadric surface, the technology of quadric surface extracting base on least square method have been implemented. The study results show that the computational accuracy of linear least square method is highest for the normal data or few noise data, and the computational accuracy of nolinear least squares method is lowest, but the computing power of the nolinear least squares method must be higher than the linear least squares method. The least squares method is sensitive to the noise data, along with the noise data increases, the precision of extracting result could come under certain influence, but the computing power is high. The research result have important guiding sense to apply the least squares method to extract quadratic surface technology for actual project application.