图像插值模糊的非线性自适应消除

针对线性移不变图像插值方法造成的插值图像边缘模糊，提出采用具有高通特性的有理滤波器增强图像，重建锐变的图像边缘。对基本有理滤波算子进行改进，避免滤波信号峰值点和孤立点的失真。推广算子的二维形式并引入图像对比度约束，形成一种自适应图像插值模糊消除算法。实验结果表明，算法能够不失真地重建图像边缘的高频细节，提高图像的视觉分辨率，图像质量优于现有图像插值模糊消除算法。

Adaptive nonlinear deblurring of image interpolation

Linear space-invariant interpolation schemes, such as the nearest interpolation, the linear interpolation and the cubic interpolation, cannot produce sufficient resolution around edges and consequently the zoomed images become obscured. Here a high-pass rational filter is improved to remove this distortion. The edge-sensitive operator is generalized to a two dimensional form to efficiently deal with images. The results of processing show that the new filter operator can preserve the geometric regularity around edges in the enlarged images interpolated by the linear space - invariant schemes. This scheme of adaptive deblurring makes enlarged images gets higher resolution and better visual quality than other image interpolation enhancement schemes, such as unsharp masking and iterative recovery.