I’m seeing 3–5% NDVI swings when fusing Sentinel-2 (10 m) with WorldView-3 pan-sharpened to 0.5 m over irrigated fields in June, even after surface reflectance and BRDF corrections. Any go-to resources or workflows for radiometric harmonization and sub-pixel co-registration that keep index stability tight, with QA I can cite?
And > radiometric harmonization and sub-pixel co-registration that keep index stability tight, with QA I’ve cut swings to about 1% by convolving WV3 to S2’s RSR/MTF and 10 m grid, then fitting a per-scene PIF-based gain/offset on red/NIR — QA is defensible by pointing to the HLS ATBD approach (Documents – Harmonized Landsat Sentinel-2); what invariant targets are you using for the PIFs?