对抗网络数字图像混合噪声滤除算法仿真

Simulation of Mixed Noise Filtering Algorithm for Digital Image of Countermeasure Network

为了提高数字图像混合噪声滤除算法的滤除效果,设计了对抗网络数字图像混合噪声滤除算法。首先对对抗网络数字图像边缘进行了检测,在虚拟现实技术下并对图像进行了预处理,将图像按照等间距形式进行采样,然后判断图像中的混合噪声,标记窗口像素,最后采用小波阈值算法对对抗网络数字图像噪声进行滤除,结合虚拟现实技术重构有效低频信号的小波系数和阈值高频信号系数,得到混合噪声滤除后的新图像,实现对抗网络数字图像混合噪声滤除。实验结果表明,经过上述研究的对抗网络数字图像混合噪声滤除算法的均方误差低,峰值信噪比与结构相似度都较高,取得了较好的应用效果。

We report a filtering algorithm against network digital image mixed noise for improving the filtering effect of digital image mixed noise filtering algorithm. Firstly, in order to sample the image in the form of equal space, the edge of the digital image was detected to preprocess the image. Secondly, the mixed noise in the image was determined to label the pixels of the window. Then, wavelet thresholding algorithm was applied to eliminate the noise of digital image. Then, based on virtual reality technology, the wavelet coefficients of effective low-frequency signal and the coefficients of threshold high-frequency signal were reconstructed to obtain a new image after mixed noise filtering. Finally, the filtering of mixed noise against network digital image was completed. The results show that the algorithm has low mean square error, high peak signal-to-noise ratio and structural similarity, and excellent application effect.