B-SPLINE SURFACE RECONSTRUCTION BASED ON FEATURE POINTS OF SECTIONAL CURVES

For given rows of data points, a new approach of B-spline surface reconstruction is presented based on feature points of sectional curves. Firstly, a new knot placement for B-spline curve reconstruction is proposed, obtaining initial B-spline curves approximating to the data points. The local curvature maximum points are recognized and extracted automatically by curvature information. The local curvature maximum points and two end points are viewed as seed points of the feature points. We construct B-spline curves approximating to the seed points, refining the feature points and updating the knot vector by B-spline curve segment complexity. The largest number of the knot vectors is regarded as candidate vectors, the common knot vectors for B-spline surface lofting can be gained by the knot tolerance and the knot insertion. Finally, the B-spine surface approximating to the data points can be accomplished by surface lofting. Compared to other B-spline surface reconstruction, our approach can reduce control points and improve efficiency of B-spline surface reconstruction.