Optimized NURBS Curve and Surface Fitting Using Simulated Annealing

Abstract

It is a key problem to approximate data of complex surface in many graphics and image processing applications. It is difficult to compute the control parameters of NURBS (Non Uniform Rational B-Spline) effecting the result of fitted shape. An evolutionary programming algorithm is used to optimize the weight and knot by minimizing the sum square error between the fitted and target curve and surface. The results obtained from curves and surfaces shows that comparing to knot optimization, the weight optimization is a better option because knot optimization requires a good initial location of knot vector.