In this paper, a new filter is proposed, in order to restore images corrupted by salt -pepper impulse noise. The algorithm consists of two main processes: 1. Making a noise position image -- an impulse detection algorithm is used before filtering, and make a noise position image; 2. Filtering -- a corrupted image is filtered by using the noise position image. If we use the noise position image, we can filter only pixels which are corrupted by impulse noise with neighbor pixels which are not corrupted by noise. In this paper, we study the first process (i.e., detection). Performance of a noise position image depends on the accuracy of noise detection. There is a two-State method as superior technique in me noise detection technique. We adopt a two -state method as noise detection, furthermore, in this paper, we introduce a multi -window method into the first process. In the multi-window method, we judge the processed point is impulse or not from all of windows which is included a processed point. We show that the PSWA filter which used the proposed method is superior to conventional methods.
展开▼