摘要： 结合混合插值样条(BIS)性质的特点，给出一类新的自适应图像插值算法。该方法可满足任意正实数倍的放缩要求，由于采用单核处理，而且插值像素点时无需求解方程组，因此计算复杂性较小。同时采用弹性边界的方法可以有效保持边缘像素原有特征。实验结果表明，该方法与三次多项式样条方法相比，具有更清晰的边界。

关键词: 图像插值, 非线性插值样条, 分段插值, 弹性边界

Abstract: This paper combines the characteristics of Blending Interpolation Spline(BIS), proposes a new method for image interpolation. The method can meet the demands of the images zoomed by arbitrary positive real number times. Using a single kernel function and without presenting additional computation in image interpolation in order to satisfy the interpolating conditions, it provides the less computational complexity. The edge-preserving characters can be realized by the flexible border. Experimental results show the method has more clear border compared with the normal cubic polynomial methods.

Key words: image interpolation, nonlinear interpolation spline, piecewise interpolation, flexible border

中图分类号: 

• TP391.41