超分辨率图像重建引起的噪声放大与滤波

总结了图像序列相位关系对超分辨率图像重建效果的影响规律,对超分辨率图像重建引起的高斯噪声和散粒噪声的放大予以研究.通过叠加不同方差的高斯噪声的图像序列重建实验,得到结论维纳滤波可以有效地滤除放大后的高斯噪声,但是超分辨率图像重建后散粒噪声放大成为"波纹"形状的噪声,传统的中值滤波法不能有效地滤除放大后的"波纹"形状噪声...

Amplificatory noise raised by super resolution image reconstruction and filter

The article summarizes the equation that phase error affects the effect of super resolution image reconstruction, researches the amplificatory noise raised by super resolution image reconstruction by test of image with noise reconstruction. There is a conclusion that Wiener Filter can also restrain the amplificatory Gauss Noise after super resolution image reconstruction, but because of Shot Noise becomes Ripple Noise after image reconstruction, Median Filter does not restrain the Ripple Noise, however, brings image confusion. In view of sequence images are processed, the article first thinks out Median Filter Based on Sequence Images, the method makes sure which pixel is with noise by pleonastic information in sequence images, and which is not, then, instead of Median Filter to the whole image, Median Filter Based On Sequence only eliminate the pixel with noise, but keeps pixels without noise still. Median Filter Based on Sequence Images restrains noise more powerfully than Median Filter based on single image, and does not bring image confusion, improves the quality of the image.