ok so ive been collecting NDVI and NDRE data with a Mavic 3M over our fields for the past three months trying to build out a proper variable rate prescription workflow. the data looks great in DJI Terra and even exported into QGIS the maps are clean. the problem is when i actually try to hand these off to our agronomist he's skeptical about the calibration and whether the reflectance values are trustworthy enough to base actual input decisions on.

like i get it, the calibration panel process is kind of janky compared to a proper MicaSense setup and the 3M doesn't have a DLS sensor so you're relying on the panel captures at the beginning and end of each flight which honestly i dont always remember to do consistently. but the spatial resolution is so much better than satellite and its way cheaper than renting a proper ag drone for monitoring only.

has anyone actually closed the loop on this and pushed Mavic 3M outputs into a real VRA prescription? or is it more of a scouting and rough zoning tool at this point

your agronomist isn't wrong to be cautious tbh. the 3M is a fantastic scouting tool and for identifying stress zones and problem areas it's hard to beat the price to value. but for actual variable rate prescriptions that are driving real input decisions i think the calibration limitations are a genuine issue not just paranoia. the lack of irradiance sensor means you're really sensitive to cloud cover changes between your panel capture and your actual flight and nobody always remembers to capture panels at both ends under consistent conditions.

that said ive seen people make it work by being super disciplined about flight conditions ? fly within an hour of solar noon, consistent cloud cover or none, always do the panel. under those constraints the data is probably good enough for zone-based VRA even if its not research-grade. whether its good enough for your specific crops and inputs is really an agronomist call.

we use it mostly for zoning yeah, like identifying the low performers versus high performers and then writing prescriptions based on those zones rather than treating the raw index values as absolute. it works pretty well for that use case. trying to use the actual NDVI numbers as direct inputs is where i think it gets shaky with the 3M calibration. but if you're just saying this area gets 20% more nitrogen than this area based on relative performance, the 3M is honestly fine for that and the res is genuinely better than most satellite stuff you can get.