Abstract: Curve reconstruction from planar unorganized points is one of the most important problems in reverse engineering. A practical B-spline curve fitting algorithm based on Goshtasby’s approach is presented. The digital image representing field distribution by estimating the bound of field strength basis function is generated at first, and then an algorithm of image thinning associated with improved BFS is proposed to overcome difficulties of obtaining the ridge contour under the situation of redundant branch pixels, and finally the weighted least squares method is used to extend the reconstructed curve for overcoming the deficiency that reconstructed curve may be shortened. Experiments show that the algorithm is valid and practical for B-spline curve reconstruction, especially when the given points are dense and noisy.

Key words: computer application, B-spline curve reconstruction, field distribution, unorganized points