一种改进的线性图像插值算法

摘要

针对传统的双线性插值法在对图像进行插值后会不可避免的产生边缘模糊的问题，提出了一种改进的线性插值法，该算法首先把待插值点分为三类，然后分别选取合适的已知点进行插值。通过对经典图像lena和pepper进行插值的实验结果表明，该算法的插值效果与双立方法相当，但计算量远远小于双立方，能有效的保持图像边缘信息，提高了图像质量。%In this paper an improved method of linear interpolation to solve the problem of the traditional bilinear interpolation which will inevitably generate edge blurring after interpolated is improved. The algorithm first converts the interpolated point is divided into three categories, and then select the appropriate known point interpolation. The experimental results of two classic images called Lena and pepper show that, the algorithm is similar with the bieubie,but the amount of calculation is much smaller than the bicubic method,it can effectively maintain the image edge information and improve the image quality.