The Application of Wavelet Transform to Satellite Cloud Image Compression

Different characteristics between meteorological satellite cloud images and common images are analyzed, and a compression method of meteorological satellite cloud images is proposed based on wavelet transform (WT). First, the bio-rthogonal Odegard 9/7 filter is selected, which shows better compression ability for cloud images. Using this filter, a cloud image is five step decomposed and reorganized. It's found that the wavelet coefficients of cloud image have more similarity and obvious attenuation in different steps, the wavelet coefficients are encoded using the improved 0 tree quantization algorithm. Finally, a further compression is processed for the output data streams using the adaptive arithmetic coder. The method proposed performs better in cloud image compression than other existing compression methods, e.g., the classic embedded 0 tree wavelet (EZW) introduced by Shapiro. And within the anamorphic error range, the highest ratio of cloud image compression is 40:1 for infrared image, 60:1 for water vapor image and 35:1 for visible image.