Comparative Analysis of Common Compression Algorithms for Sentinel-2 Multispectral Band Compression

Sentinel-2 Earth observations generate massive volumes of imagery data stored in the JPEG 2000 file format. The paper evaluates lossless compression strategies for satellite data by comparing two general-purpose methods (Deflate, Bzip2) and two image compression algorithms (JPEG 2000, JPEG XL). The experiments focused on individual band compression. Parallel processing implementations were tested, to leverage the features of the selected algorithms. The experiments were conducted on a dataset consisting of full tile Sentinel-2 L2A images with reduced cloud levels, where two performance metrics were measured, a compression ratio and a throughput.