摘要
针对传统变换单幅图像处理方法缺少提取不同位置信息的步骤,导致边缘保持不完整、运算时间长等问题,提出基于互补色小波变换的单幅图像边缘保持方法。对图像加入高斯白噪声,运用软阈值手段实现单幅图像去噪。构建互补色小波分区归一化变换模型,将互补色小波子带与图像内每个方向采取等权重融合,使用三个小波层级尺度提取不同位置的图像信息。运用边缘保持局部自适应图像插值法,根据自适应临界值Canny边缘检测,将子图像区分为平坦区域和边缘区域,并依次进行双线性插值和非线性插值处理,保证单幅图像边缘保持完整性。仿真结果证明,研究方法的图像边缘保持效果更佳,运算时间更短,具备很好的应用价值。
In traditional image processing methods, the error of edge information recognition is large and the operation time is long. Therefore, a single image edge preservation method based on complementary color wavelet transform was put forward. Firstly, white Gaussian noise was added to the image, and then the soft threshold was used to remove the noise from a single image. Secondly, the complementary color wavelet partition normalization transformation model was constructed, and the complementary color wavelet sub-band was integrated with each direction of image by equal weights. Thirdly, three wavelet hierarchical scales were used to extract the image information at different locations. Moreover, local adaptive image interpolation method based on edge preservation and Canny edge detection based on adaptive critical value were used to divide the sub-image into flat region and edge region. After that, bilinear interpolation and nonlinear interpolation were carried out successively. Finally, the edge integrity of a single image was remained. Simulation results show that the proposed method has better image edge preserving effect, shorter operation time and higher application value.
作者
杨柳
王望
王兴
YANG Liu;WANG Wang;WANG Xing(Henan University of Economics and Law,Zhengzhou Henan 450006,China;Taiyuan University of Science and Technology College of Computer Science and Technology,Taiyuan Shanxi 030024,China)
出处
《计算机仿真》
北大核心
2020年第5期477-481,共5页
Computer Simulation
关键词
互补色小波
图像去噪
边缘保持
自适应
Complementary color wavelet
Image denoising
Edge preservation
Adaptive
