Efficient Data Compression Techniques for Weather Data

摘要

We collaborated with the Naval Research Laboratory (NRL), Marine Meteorology Division in Monterey, CA to develop novel and efficient data compression techniques for weather data. We developed a differential coding algorithm for lossless compression of radar data by using motion estimation and compensation techniques, with the collateral development of a fast block matching method to improve the computational efficiency of motion estimation schemes. We also improved the compression efficiency of the existing UF file compressor by introducing a new coding method, which exploited the large concentration of blankout markers in the quality-controlled radar data. Furthermore, we studied the feasibility of further compression of the satellite weather images stored in JPEG format. As a result of collaboration with the NRL researchers, the improved and standardized UF radar data file compressor software has been transitioned to testing and installation on ships.