【摘要】 为了解决传统插值方法造成图像模糊、锯齿以及边缘细节信息缺失的问题,提出了一种新的边缘图像插值算法。该算法通过待插值像素点周围低分辨率像素之间位置和方向的特征计算其标准差并将其作为局部动态阈值,以此作为划分图像区域的比较阈值,将图像划分非边缘区域和边缘区域;然后对非边缘区域进行双线性插值,边缘区域进行自适应插值。实验结果表明,与传统的插值方法相比,该插值方法较好的保持了图像边缘细节信息,且具有较好的视觉效果。更多还原

【Abstract】 In order to solve blurring,sawtooth and edge details missing for image caused by,the traditional interpolation methods a new edge image interpolation algorithm is proposed. The proposed algorithm calculates the standard deviation as local dynamic threshold by the relationship of position and direction around the low resolution pixels,which is regarded as a comparison of threshold of dividing image into non-edge area and edge area. Then bilinear interpolation on the non-edge area,and adaptive interpolation on the edge area. The experimental results show that the interpolation method is better to keep the image edge details,and has a good visual effect compared with traditional interpolation method.更多还原