Abstract: A new adaptive method of image de-noising based on two-dimensional wavelet transform and independent component analysis(ICA) is presented.It does not need a lot of observed image samples,and it is not necessary to know the details of the observed image signal type in advance.A single observed image could be de-noised by this method adaptively.The principle of ICA virtual observed noise channels de-noising and two-dimensional wavelet transform constructing those channels are introduced,and the experimental results based on this method are given.The results show that by using the proposed method the hybrid noise is removed effectively and the image details are protruded and the image quality is improved.