Data Fitting by Quadratic Splines

In this paper the problem of data fitting using quadratic spline is addressed. An efficient algorithm is presented. Given a set of ordered planar points, the algorithm divides this set of points into subsets, while the data points in each subset are collinear within the given tolerance. The data points in each subset are fitted by a curve segment and all data points are fitted by a C¹ quadratic spline curve. The algorithm is simple and reliable, decreases the number of fitting curve segments and maintains the approximating accuracy. This algorithm is tested and can be applied to reverse engineering and image curves fitting.