The goal of this paper is to develop a fast noise reduction (FNR) approach to infrared (IR) images. The IR images usually have noise, edges, text information and small objects of interest. The well known standard median filter (MF) is not suitable to process these IR images. The main characteristics of the proposed approach are that no prior knowledge about the IR image noise is necessary and no parameters must be preset. It is developed based on IR imaging mechanism to detect the noisy pixels and on modified median-based filter to remove the noisy pixels in IR images. The FSN approach has the advantage of low computation load. In addition, it can retain reasonably good edges, details and texture information even if the size of the filter window increases. The goal of the proposed approach is feasible for an effective process for reducing noise of IR images, regardless of the filter windows' and objects' sizes. It also significantly speeds up the noise reduction process and can apply to real-time video processing. FNR approach will extend the IR image applications of medicine. (C) 2010 Elsevier B.V. All rights reserved.
