DifferenceCompositor

[djamainajib]

Describe the bug
A clear and concise description of what the bug is.

To Reproduce

# Your code here

Expected behavior
A clear and concise description of what you expected to happen.

Actual results
Text output of actual results or error messages including full tracebacks if applicable.

Screenshots
If applicable, add screenshots to help explain your problem.

Environment Info:

• OS: [e.g. OSX, Windows, Linux]
• Satpy Version: [e.g. 0.9.0]
• PyResample Version:
• Readers and writers dependencies (when relevant): [run from satpy.utils import check_satpy; check_satpy()]

Additional context
Add any other context about the problem here.

[gerritholl]

What is your question, or what is the problem you are facing? When filing a bug report, you are supposed to replace the boilerplate text with specific descriptions.

[djamainajib]

Issue: Large discrepancy between diff_m15_m12 composite and separately computed difference m15 - m12
Context:
Using polar2grid, I generated three products from VIIRS data for the same area and time:

• m15
• m12
• diff_m15_m12 (difference product computed using the DifferenceCompositor)

The compositor configuration for diff_m15_m12 is as follows:

diff_m15_m12:
  compositor: !!python/name:satpy.composites.DifferenceCompositor
  prerequisites:
    - name: M15
    - name: M12
  standard_name: M15-M12

Problem:
When I compare the diff_m15_m12 product to a difference image I compute manually as m15 - m12 (subtracting the calibrated m12 product from the calibrated m15 product), I observe a large discrepancy between the two results (Figure bellow).

Additional details:

• The images are in 0–255 format.
• I calibrated each product by applying the offset and scale factors provided in the metadata of each polar2grid output before performing the difference computation manually.

Questions:

• Is the DifferenceCompositor correctly applying calibration and offset internally before difference calculation?
• Any suggestions on how to debug or verify the internal compositor calculations?

Thank you!

[djhoese]

The difference is most likely due to the enhancement on the compositor. Did you configure an enhancement in Polar2Grid? You mentioned scale factor and offset, but can you provide the actual code you're running. And if you have it, could you attach the Polar2Grid log file?