Adaptive switching median filter for removal of impulse noise

To overcome the drawbacks of the traditional median filter, the adaptive switching median (ASM) filter is proposed. In the algorithm, the two-stage noise detection scheme is adopted to classify the pixels in the image into signal pixels and noise pixels. The number of noise pixels is computed to estimate the noise density, based on which the filtering means is chosen adaptively to remove impulse noise Simulation results demonstrate that the ASM filter can suppress noise very effectively while preserving fine details very well and has excellent filtering performance.