In image processing, a natural and normal action is to capture an image. In capturing an image, interfering noise from certain sources contaminates the image and degrades its quality. This noise may not be erased easily. Nonlinear filters can achieve noiseless pictures better than linear ones. In this article, an effective and robust method against interfering noise in restoration of infrared images is proposed. An improved hybrid median filter is initially introduced in this proposal, and in conjunction with a Wiener filter the noise is removed using a method based on wavelet transform. Combination of spatial and wavelet transform domain filters is the idea proposed in this paper to cover the shortcomings of both domains. The results of simulations show that the proposed methods are better than other popular noise rejection methods in terms of image quality and PSNR.