An Adaptive Denoising and Enhancing Algorithm Based on the MAP Rule in the Contourlet Domain for Infrared Image

In order to solve the problem of weakening the details and the edges of image while denoising in the contourlet domain, this paper presents an adaptive denoising algorithm with detail enhancement and applies it to the denoising procedure of infrared image. On the basis of the assumption that the prior distribution of the original image coefficients and the noise´s are both Gaussian in the contourlet domain, this method firstly makes use of the rule of Maximum a Posteriori to compute the shrinkable factor for the contourlet coefficients, then modifies it by taking decomposable scale and directional energy into account. Finally, a denoised and enhanced image can be obtained after the contourlet coefficients, which are shrunk by the modified shrinkable factor, are made by the reverse contourlet transform. The contourlet transform has not the character of translation-invariance, so the cycle spinning method is applied to the whole denoising procedure to overcome the drawback. The experimental results show that the method given by this paper, compared with the general denoising algorithm of wavelet and contourlet, can enhance image details, stretch image contrast and produce a good visual effect though it has a little loss of PSNR (Peak Signal Noise Ratio). The enhancing idea of coefficients in the contourlet domain proposed by this paper can apply to other algorithms based on proportional shrinkage.