Maya4 is a project dedicated to curating and providing multi-level intermediate SAR representations from Sentinel-1 acquisitions, spanning the entire chain from Level 0 to Level 1.
The name Maya4 draws inspiration from the MΔyΔ veil in philosophy, where reality is hidden behind successive layersβjust as radar echoes undergo transformations before forming a final SAR image.
Maya4 provides Pre-Training (PT) dataset exposing intermediate SAR signal states in a modern (Zarr Format):
π Dataset Split | π Contents | π°οΈ Acquisition Mode | πΎ Size |
---|---|---|---|
pt1 | LINK | Stripmap | 17 TB |
pt2 | LINK | Stripmap | 17 TB |
pt3 | LINK | Stripmap | 17 TB |
pt4 | LINK | Stripmap | 17 TB |
Total | β | β | 68 TB |
A major characteristic of this dataset lies in its sharding multi-level organisation:
π Level | π€ Abbrev. | π Description | π― Purpose / Value |
---|---|---|---|
π‘ Raw | raw | Unprocessed radar echoes as recorded by Sentinel-1 | Baseline data; allows full custom SAR processing |
ποΈ Range Compressed | rc | Echoes compressed in the range dimension via matched filtering | Improves SNR; isolates scatterers along range |
π― Range Cell Migration Corrected | rcmc | Motion-compensated echoes with corrected range migration | Preserves geometric fidelity; enables proper azimuth focusing |
πΌοΈ Azimuth Compressed | ac | Fully focused SAR image in slant-range geometry | Standard Level-1 product; interpretable imagery |
The following use cases can be addressed with Maya4: