Digital Data Compression Using Vector Quantization and Double Color SpaceNormalization

摘要

The Naval oceanographic and Atmospheric Research Laboratory's (NOARL) Map DataFormatting Facility (MDFF) is developing improved methods of compressing full-color (24-bit) aeronautical chart images. The original MDFF compression method consists of two stages. In the first stage, color compression, the 240 most representative colors in an image are identified and entered in an 8-bit color lookup table. Each 24-bit pixel (8 bits each of red, green, and blue) in the image is then replaced with one of the 8-bit values in the table, for a compression of 3:1. In the second stage, the image is spatially compressed by replacing each rx)n-overlapping square of 2x2 pixels with a 1-byte codeword, for an additional 4:1 compression. A tote( of 256 codewords are stored with their four corresponding pixel values in a codebook file. This paper describes a new method In which 4x4 pixel sets (rather than the original 2x2 pixel sets) are replaced by codewords in the spatial compression stage, achieving an 8:1 compression. The first stage of color compression remains the same. This new method achieves a total compression ratio of 48:1 (3:1 color and 16:1 spatial) when using 8-bit codebook lookup values or 24:1 (3:1 color and 8:1 spatial) when using 16-bit codebook lookup values. Digital maps, Data bases.