Astronomical Image Compression Based on Noise Suppression

We review methods which are widely used for compression of astronomical images. These include multiresolution approaches. We use such an approach, a pyramidal median transform, which is followed by noise suppression at each resolution scale. To preserve photometric and astrometric fidelity to the input image, a multiresolution astronomical-object support is used. We discuss the common astronomical noise models (additive Poisson, Gaussian, and Poisson plus Gaussian). A range of examples on digitized plate, and on CCD, image data is used to show the effectiveness of this approach.