Compression experiments with AVHRR data

This paper describes several data compression approaches for producing browse data from AVHRR data, and evaluates these approaches qualitatively and quantitatively. They include a hierarchical data compression scheme based on progressively finer image segmentations, and various vector quantization approaches. They are evaluated in terms of compression ratio (or data rate), computational requirements, and the image and analysis errors introduced due to lossy compression. Analysis products evaluated for error include cloud coverage area and sea surface temperatures.>