A technique for lossless compression of GMS VISSR images

In order to analyze weather conditions, it is useful to get VISSR images of GMS with high brightness resolution. As the data volume is very large, data compression is indispensable to transmit and/or store them. The authors proposed an effective approach for lossless compression of them. The value at each pixel is predicted from the values of pixels around it by means of a linear predictor in the first stage. Then, the error value is obtained as the difference between the real value and predicted one. The error value is again predicted from error values of pixels around it by means of a two-dimensional nonlinear predictor and a residue is obtained as the difference between the error value and prediction one in the second stage. The residues are encoded by Huffman method in the final stage. The data volume of VISSR image can be reduced to 32% by this method. The value is about 20% less than that obtained by the spatial method of JPEG