Abstract:Highlights•We introduce a new method to interpolate scattered 3D data points with a single B-spline surface patch, whose knot vectors are generated according to the parametrization of the scattered data points by iteratively inserting knots to the knot vectors. The generated knot vectors guarantee the existence of interpolation surfaces, while often using fewer control points.•Our method introduces a novel parameter δ to improve the stability of control point solving. By increasing δ, the interpolation errors will be reduced.•Since our knot vector construction algorithm guarantees the existence of the interpolation surface, the control points can be solved by a global fairing energy minimization procedure with strict interpolation constraints. Thus, a globally smooth B-spline surface can be generated.We introduce a new method to interpolate scattered 3D data points with a single B-spline surface patch which is globally smooth. Given a set of scattered 3D data points and their corresponding parametrization, our method first constructs a set α of B-spline bases using a weighted strategy, and inserts knots to the knot vectors based on α. Then, the knot insertion procedure is iterated until a set β of B-spline bases exists, which indicates the existence of the interpolation surface. Finally, by applying the fairing energy minimizing with interpolation constraints, a globally smooth B-spline surface which interpolates the data points can be produced. Experimental results demonstrate that the generated B-spline surfaces often have fewer control points than those of traditional methods, while keeping the scattered data points interpolated accurately. The implementation of our algorithm and the scripts to reproduce all the results are available at https://github.com/wangbolun300/sparse_data.

Nonparametric Density Estimation with Nonuniform B-spline Bases

Adaptive Nonparametric Density Estimation with B-Spline Bases

A local fitting algorithm for converting planar curves to B-splines

Uniform Trigonometric Polynomial B-Spline Curves

Nonparametric Density Estimation for Data Scattered on Irregular Spatial Domains: A Likelihood-Based Approach Using Bivariate Penalized Spline Smoothing

Kernel density regression in the additive model: a B-spline approach

B-spline surface fitting by iterative geometric interpolation/approximation algorithms

Efficient spline orthogonal basis for representation of density functions

Distribution and density estimation based on variation-diminishing spline approximation

The Unimodality of Initial B-Spline Approximations in Spline Fitting

Efficient Basis Selection for Smoothing Splines Via Rotated Lattices

Fast Bivariate Penalized Splines: the Sandwich Smoother

On A Bivariate B-Spline Basis

More efficient approximation of smoothing splines via space-filling basis selection

A Hybrid Parametrization Method for B‐Spline Curve Interpolation via Supervised Learning

Scattered Points Interpolation with Globally Smooth B-Spline Surface using Iterative Knot Insertion

Adaptive Bayesian Multivariate Spline Knot Inference with Prior Specifications on Model Complexity

Empirical Density Estimation based on Spline Quasi-Interpolation with applications to Copulas clustering modeling

EM Estimation of the B-spline Copula with Penalized Log-Likelihood Function

The Bases of the Non-Uniform Cubic Spline Space

Convergence of Geometric Interpolation Using Uniform B-splines.