一种反梯度自适应插值算法

Inverse-gradient adaptive interpolation algorithm

针对边缘方向插值算法计算量大、边界的处理效果不理想,而自适应放大算法对权值因子的选择太过简单的问题,提出了一种新算法。根据数字图像边缘区域的灰度值变化强烈,而非边缘区域的灰度值变化比较平滑的特点,在边缘处沿边缘方向作线性插值,非边缘区域则采用反梯度自适应插值的方法。试验结果表明,此算法不仅能降低运算复杂度,还能获得清晰的图像边缘,进一步提高了插值图像的视觉质量。

In view of the inherent shortcomings of edge-directed interpolation algorithm and locally- adaptive zooming algorithm, a new algorithm is proposed. The characteristic of the digital image is that the gray value in edge area is sharp, and the gray value in the non-edge area is smooth. In accordance with this characteristic, the linear interpolation was used along the edge orientation, and inverse gradient interpolation was used in the non-edge area. Experimental results demonstrate that this algorithm not only decreases the computational complexity, but also obtains the clear edge image and substantially improves the visual quality of the interpolated images over conventional linear interpolation.