Analytic processing of experimental data by a B-spline fitting

The present paper reports a bi-dimensional fitting procedure based on the use of B-spline functions applicable to differently shaped domains. The analytical model is based on a number piecewise polynomial functions defined in a limited portion of the total domain. The number of the functions and the order of the polynomials can be chosen independently from each other; in this way data with a complex spatial distribution can be fitted simply by fixing the order of polynomials and increasing the number of functions. The proposed approach was applied to numerical generated data. By the numerical analyses a procedure to evaluate the number and the order of polynomials was defined, and the capability of the approach to be noise tolerant was estimated.