WorldView-3 VNIR tiles with SWIR swaths

[oceancolorcoder]

Hi Quinten,

After batching a couple hundred WV-3 files in recently, I realized that occasionally sequential images are defined (and named) for the same datetime to the second, and yet represent different tiles of a swath. In this case, my batch sees the acolite output files, and skips the second file if it’s already written or overwrites it with the new tile. I poked around a bit in acolite, but did not see how you deal with two or more tiles from a swath with the same datetime, e.g., by adding the Digitalglobe sequential P00x number to the output file name to distinguish sequential tiles in a swath.

An example is tiles:
20JUL25154338-M1BS-505952035100_01_P001
20JUL25154338-M1BS-505952035100_01_P002
which correspond to SWIR swath:
20JUL25154338-A1BS-505952043100_01_P001

Just wondering if you’d encountered this before and had considered how to work around it. It looks like the ofile/oname is defined in l1_convert prior to splitting off the tile (or “P”) number, and then I’m foggy on whether the tile captured is actually picking off the tile number from the SWIR “A” file or the VNIR “M” file. They are not the same in many cases: I often receive the data with one long swath of a SWIR images corresponding to several VNIR tiles. The tile number in the name of the result should be the one associate with the VNIR file.

Cheers,
Dirk

[quinten]

Hi Dirk

I haven't used the WorldView processing all that much myself. The files I have with different tiles are in the same bundle.

Perhaps it is worth using the region_name= setting to add an additional identifier to the output filename? E.g. add the tile name as region_name?

Quinten