Submit Manuscript  

Article Details


A Novel Adaptive GA-based B-spline Curve Interpolation Method

[ Vol. 13 , Issue. 3 ]

Author(s):

Maozhen Shao, Liangchen Hu, Huahao Shou* and Jie Shen   Pages 289 - 304 ( 16 )

Abstract:


Background: Curve interpolation is very important in engineering such as computer aided design, image analysis and NC machining. Many patents on curve interpolation have been invented.

Objective: Since different knot vector configuration and data point parameterization can generate different shapes of an interpolated B-spline curve, the goal of this paper is to propose a novel adaptive genetic algorithm (GA) based interpolation method of B-spline curve.

Methods: Relying on geometric features owned by the data points and the idea of genetic algorithm which liberalizes the knots of B-spline curve and the data point parameters, a new interpolation method of B-spline curve is proposed. In addition, the constraint of a tangent vector is also added to ensure that the obtained B-spline curve can approximately satisfy the tangential constraint while ensuring strict interpolation.

Results: Compared with the traditional method, this method realizes the adaptive knot vector selection and data point parameterization. Therefore, the interpolation result was better than the traditional method to some extent, and the obtained curve was more natural.

Conclusion: The proposed method is effective for the curve reconstruction of any scanned data point set under tangent constraints. Meanwhile, this paper put forward a kind of tangent calculation method of discrete data points, where users can also set the tangent of each data point in order to get more perfect interpolation results.

Keywords:

B-spline curve interpolation, genetic algorithm, tangent vector, tangent constraints, discrete data, knot vector.

Affiliation:

College of Science, Zhejiang University of Technology, Hangzhou, College of Science, Zhejiang University of Technology, Hangzhou, College of Science, Zhejiang University of Technology, Hangzhou, Department of Computer & Information Science, University of Michigan-Dearborn, Dearborn, MI

Graphical Abstract:



Read Full-Text article