图像放大中的边缘细化算法研究

ON EDGE THINNING ALGORITHM IN IMAGE ZOOM

当图像放大时边缘会出现锯齿失真,为了抑制这种失真,提出了边缘阶梯细化的图像放大算法。该算法在图像的非边缘区域采用经典的双线性插值算法;同时根据Canny边缘检测的结果,进一步进行阶梯检测,滤除不会产生锯齿失真的竖直和水平边缘;而在使用经典算法会产生严重锯齿失真的边缘区域,运用基于阶梯细化的插值算法;该插值算法在放大图像的同时,抑制边缘中的阶梯被放大,从而达到减少锯齿失真的目的。实验结果表明,该算法比经典算法更能有效地抑制锯齿失真。

Jaggies distortion on the edges may occur in image zooming.In order to restrain such distortion,in this paper an image zooming algorithm with edge stairway thinning is proposed.The algorithm adopts classical bilinear interpolation algorithm in the area of image other than the edges;concurrently the stairway detection is further made according to the outcomes of Canny edge detection to sift out the horizontal and vertical edges which will not produce jaggies distortion;but in edge areas where severe jaggies distortion will occur when using classical algorithm,the stairway thinning-based interpolation algorithm is implemented.This interpolation algorithm keeps stairways from being enlarged when zooming out the image,hence the goal of reducing jaggies distortion is achieved.Experiment results show that this algorithm can restrain jaggies distortion more effectively than the classical algorithm.