Abstract: Image enlargement and reduction are widely applied in real life. The popular image processing software commonly carries out magnification and reduction by interpolation methods. The realization of a variety of interpolation algorithms is now a hotspot for research. Now a new interpolation algorithm based on the nonuniform Bspline curve was proposed. By this way which applies a nonuniform parameterization method different from the traditional algorithms, the objective image after interpolation can maintain a higher resolution and good smoothness. As for the serrated phenomenon on image edges, the algorithm interpolates the image edges with the bi-cubic algorithm when interpolating. Thus, the visual effect on edges can be optimized. The experimental results show that the image becomes clearer and smoother by the new way. On the image edges, the serrated phenomenon is eliminated and a good visual effect is achieved.

Key words: image enlargement, B-spline curve, edge check, bi-cubic interpolation

摘要： 图像放大和缩小处理在实际生活中具有广泛的应用，常用的图像处理软件普遍采用插值方法进行放大和缩小，各种插值算法的实现是目前研究的热点。提出了基于三次B样条函数的插值算法，采用不同于传统算法的非均匀参数化方法，使得插值后的图像能够保持较高的清晰度和平滑度。针对图像边缘处的锯齿现象，该算法在图像插值时对边缘像素采用双三次插值，优化了图像边缘的视觉效果。实验结果表明，使用该算法放大后的图像平滑清晰，消除了图像边缘处的锯齿效应，取得了良好的效果。

关键词: 图像放大, B样条曲线, 边缘检测, 双三次插值