基于梯度最小准则的线性插值修正算法

针对传统的线性插值存在的图像边缘模糊问题，提出了一种改进方法。根据边缘与梯度之间的关系,以插值点梯度最小为准则,对线性插值结果进行修正,实现保护插值图像的边缘信息。实验结果表明，该方法与传统的线性插值法相比，插值图像的平均梯度提高,均方误差减小，是一种有效提高插值效果的方法。

Linear interpolation correcting algorithm based on gradient minimization

To solve the image edge blurring which existing in traditional linear interpolation,an improved method has been proposed.According to the relation of edge and gradient,the correcting to linear interpolation,based on gradient minimization,protects the edge information of the interpolated image.Experiment results show that compared with traditional linear interpolation,this method increases the average gradient,while decreases MSE of the interpolated image,which effectively improve the quality of the interpolated image.