The impact of lossy compression on the output of supervised and unsupervised classification products

The quantity of data generated by remote sensing platforms is increasing at a significant rate. Several countries have now their own remote sensing satellites, and commercial groups plan to have sensors that provide data to a wider market. In addition to the growing number of data providers, technological advances resulted in new remote sensing systems that generate data with higher spectral, spatial and radiometric resolutions that translate into larger data sets. Although some of these developments are justified by improved results in special purpose remote sensing applications, they contribute to the data management problem. This paper analyzes the nature of the information base included in remote sensing imagery and the impact of compression techniques on the effectiveness in user applications