search

nasa-nccs-hpda / evhr

Enhanced Very-high Resolution (EVHR)
4 stars 0 forks source link

EVHR Notes for Next Release #15

Open jordancaraballo opened 1 year ago

jordancaraballo commented 1 year ago

Using this issue to document EVHR next release suggestions/requirements.

mwooten3 commented 11 months ago

EvhrToa:

  1. Updated TOA coefficients

  2. Abandon status filter in query and/or enable option to check file existence on disk instead

  3. Address duplicate file issue that still exists

  4. Option to check for final toa output before running strip to skip unneeded processing

  5. Option/method to check if strip is being processed elsewhere to enable processing same list of outputs on multiple nodes concurrently (currently looking in 1-bands dir but this can be messy e.g. if processing gets hung/restarted)

  6. Pansharpen updates: 6a. Potentially re-evaluate pansharpen scene selection process (IIRC current process runs all scenes in a strip rather than only those matching the MS scenes being processed) 6b. Potentially re-evaluate the way pansharpen process finds scenes (e.g. enable option to supply pan file paths to command, bypassing database query, to avoid having to rely/wait on archive ingest) 6c. Potential issue with pansharpen process not skipping if processing error (e.g. when duplicates are included). Regular ToA process does this, pansharpen process should as well 6d. Output file renaming:

    • Pan outputs should have P1BS rather than M1BS in filenames
    • Temporary pan output should be renamed ("*M1BS_*-toaForPanSharp.tif" to something like e.g. "*P1BS*-toaTemp.tif") and/or deleted after warp step
    • Final pan output could be renamed ("*M1BS*-warpedPsToa.tif" to something like e.g. "*P1BS*-toa.tif" or toa for pansharp), with option to keep pan outputs
    • .xml files (both P1BS and M1BS) should be copied/renamed to exist beside all possible final outputs (i.e. *M1BS*toa-sharpened.tif; *P1BS*toa or *toaForPansharp.tif, whatever), not just for *M1BS*toa.tif. However, consideration maybe given to the fact that some items in these .xml's do not accurately reflect their .tif partners (which is also true of the current .xml's that go with the standard 2m M1BS toa.tif outputs to a lesser extent)