基于方向变换和方向波的图像去噪

Image Denoising Based on Direction Transformation and Directionlets

针对小波变换处理二维信号时存在方向缺失、不能达到最优逼近的缺点,借鉴方向波中处理边界问题的思想,利用方向滤波器组中的采样矩阵实现图像旋转。将具有一定夹角的2个方向利用采样矩阵表示,使这2个方向分别旋转投影成为水平方向和垂直方向。考虑到小波分析擅长处理纹理特征的优点,结合方向变换与小波变换,将对角线信息进行优化表示。仿真实验验证该方法可显著改善图像的视觉效果。

Aiming at the disadvantages of wavelets, which exist the absence of directions and cannot achieve the best approximation when deal with 2D signals. The idea of processing edge problem in directionlets is introduced. Sample matrix in directional filter banks is used to perform the rotation for images and presentation for two directions with angular, which makes the two directions project be vertical and horizontal direction. At the same time considering that wavelets are good at dealing with texture features, combine the direction transformation with wavelet transform further. The better presentation for diagonal information is achieved. Simulation experimental results show that the method can significantly improve the visual effect of the image.