摘要
传统插值算法容易产生细节模糊和边缘锯齿效应,而边缘是图像的重要信息,直接影响插值图像的视觉效果。为了解决这一问题,提出了一种新的基于误差修正的自适应边缘保持插值算法。首先将原图像向插值图像进行映射,再定义一个3×3的滑动窗口,用水平和垂直两个模板确定边缘类型,根据边缘类型进行插值,最后对仍未定义像素点用插值误差定理进行处理,总体上对所有像素都进行了分情况插值处理。实验结果表明,本算法有效地保持了图像的边缘信息,能够获得视觉质量较好的插值图像,且算法比较简单,容易实现。
Conventional interpolation algorithms suffer visually the effects of blurred details and jagged artifacts in image to some extent.Edges are important to an image and have direct effect on interpolation images.To solve these problems,a new algorithm for image interpolation based on error correction was proposed.Firstly,map the original image to the interpolation image.Then define a 3×3 sliding window and decide the type of edges using a horizontal and a vertical edge masks,interpolate the image according to the types of edges.Finally,evaluate the undefined pixels using interpolation error theorem.Experimental results show that the proposed method effectively maintains the edge information and can obtain better visual quality of interpolated images.The algorithm is simple and easy to implement.
出处
《计算机技术与发展》
2011年第2期116-119,共4页
Computer Technology and Development
关键词
图像插值
边缘
误差修正
image interpolation
edge
error correction
