Aiming at the characteristic of the actual image, which is low contrast, complex background and the high background noise, a new image denoising method based on liftwavelet analysis and median filter technology is proposed. Firstly, the noise image is decomposed with the liftwavelet. Second, the high frequency parts of decomposed image are carried on median filter algorithm to improve the removing result of the noise image. The denoising image is obtained to reconstruct the high frequency parts processed and low frequency parts of decomposed image. Finally, the image signal to noise ratio (SNR) and the rootmeansquare error (RMSE) and the image gray surface chart are applied to estimate the denoising effect of the nearinfrared images. These removing noise methods, such as the ordinary wavelet filter, the median filter and so on, are applied to remove the image noises. The experimental results indicate that this method both can eliminate the actual image noise and maintain image edge information. It can remove effectively noise of the real images.