Earth observation data provenance protection through self-recalibrated watermarking

Earth observation (EO) data are the foundation for any web service relying on remote sensing and geographic information systems. Protecting their provenance is crucial yet challenging, given their susceptibility to rapid and widespread copying and dissemination. Often, when tampering or unauthorized use of the data is detected, the data has already been disseminated. This work presents an effective watermarking technique for tracking EO data copies and validating their provenance and authenticity without compromising remote sensing algorithms’ functionality. Our approach allows the relocation of the watermark while ensuring the operability of the database. This feature is particularly useful when the database content’s priorities and attributes’ tolerance to distortion change over time. Experimental results indicate that our method outperforms existing techniques regarding data provenance and tampering detection when considering multi-type digital assets repositories. It also establishes the foundation for protecting linked content stored using different data types.