An Image denoising method combining dt-cwt and Fourth-order PDE

It is proved that image gray level tends to be piecewise constant in the low order partial differ Ential equation denoising model. Considering the applications of wavelet multi-scale decomposition in image processing, a kind of remote sensing image Denoising model combining dual tree complex wavelet transform and fouth order partial differential equation are put forward. Firstly, the noise images are multiscalely decomposted by DT-CWT. Then The low frequency components are reserved and noise in high frequency components of other layer are removed by th E Fourth order PDE model. Finally, the high and low frequency components of the the corresponding layer are reconstructed the final IMA Ge. According to the denoising experiments results of the ZY-3 satellite images which have different noise intensities proved that's average PSNR of results by proposed model increases 1~2 DB and the structural similarity increases as. The denoising model proposed can effectively preserve the image details while removing noise.