Experiments in the lossless compression of time series satellite images using multispectral image compression techniques

This paper is a preliminary study reporting on the results of experiments to take advantage of interband correlation between images from the GOES-West satellite. Experiments primarily focus on attempting to improve the prediction of pixel intensities using information between images in a close time sequence. We apply the use of motion vectors, linear and nonlinear predictors, as well as context modelling. Initial results show some promise, but more study is needed.