Reconstruction is the analysis service that estimates an image or a set of data values from other measurements when the quantity of interest was not recorded directly.12

SWEET declares Reconstruct equivalent to Reconstruction and places it under DataAnalysisService, without definition text.3 In nuclear medicine tomography, a Journal of Imaging article states that the reconstruction task is an inverse problem in which the underlying cause must be estimated from the observed data.1 That is a nuclear medicine imaging sense, retained here only to show the inverse-problem meaning of the word.

In Earth observation the nearest use is filling cloud-induced gaps. A Scientific Reports study begins from the point that cloud cover creates frequent data gaps in Sentinel-2 imagery, and it tests gap-filling methods on simulated cloud scenarios.2 In that study the spatio-temporal category, especially clustering and linear regression, showed the highest accuracy and robustness across the gap scenarios, which is a result for one evaluation and not a ranking for all sensors.2

A second class of the same name sits in the provenance module, as a subclass of ExperimentActivity.4 Retained sources do not establish that the two are the same concept, or a single agreed method.

Footnotes

  1. Journal of Imaging (MDPI) via Europe PMC, “Iterative Reconstruction with Dynamic ElasticNet Regularization for Nuclear Medicine Imaging”. https://www.ebi.ac.uk/europepmc/webservices/rest/PMC12294920/fullTextXML ↩ ↩2

  2. Scientific Reports (Nature Portfolio) via Europe PMC, “Evaluating Sentinel-2 gap filling techniques for cloud removal and data reconstruction”. https://www.ebi.ac.uk/europepmc/webservices/rest/PMC13004938/fullTextXML ↩ ↩2 ↩3

  3. ESIP SWEET ontology (GitHub mirror), “reprDataServiceAnalysis.ttl”. https://raw.githubusercontent.com/ESIPFed/sweet/master/src/reprDataServiceAnalysis.ttl ↩

  4. ESIP SWEET ontology (GitHub mirror), “reprSciProvenance.ttl”. https://raw.githubusercontent.com/ESIPFed/sweet/master/src/reprSciProvenance.ttl ↩