okay so this has been bugging me for a while. running a Mavic 3 Multispectral for NDVI maps on a client's wheat fields and getting decent data but i keep second-guessing whether the multispectral is actually catching stress earlier than a good RGB setup with the right processing pipeline would. like theoretically yes obviously NDVI is doing more than just visible light but in practice im not sure my clients can even act on the data fast enough to make the early detection matter.

part of the problem is the flight planning ? to get good multispectral data i need consistent lighting conditions and my schedule is at the mercy of cloud cover and client availability and half the time by the time i fly the stress is visible anyway. a guy i know just uses a Phantom 4 Pro with RGB and does very frequent flights like every 5-6 days and swears his change detection workflow catches problems just as early.

is the multispectral actually earning its keep here or is flight frequency and consistency just more important than sensor type

this is genuinely a good question and i dont think there's a clean answer. from what ive seen doing this for a few years the multispectral advantage is most real for detecting stress before visible symptoms show up ? we're talking like 7-14 days ahead depending on the stressor. but you're right that the operational overhead is real. sun angle, calibration panel, consistent altitude and overlap, it all adds up and if you're not flying under good conditions your NDVI maps are kind of garbage anyway.

the frequent RGB flight approach your contact does is underrated honestly. change detection on well-processed RGB orthos can catch a lot, especially things like waterlogging or nitrogen deficiency that show up pretty clearly in canopy color before they look bad to the naked eye from ground level. where i think multispectral actually justifies itself is for clients who are making variable rate application decisions and need the zone maps to hold up agronomically ? an agronomist looking at those maps needs the spectral data to be confident recommending a prescription, RGB change detection is a bit harder to tie to an action threshold.

so anyway i had almost this exact debate with a precision ag consultant last spring and his take was basically that for most row crop clients the bottleneck isn't detection it's response time and decision making. like even if you catch stress 10 days earlier with multispectral if the farmer cant get a spray rig scheduled for two weeks it doesnt matter. so he said optimize for whatever gets you flying more consistently and reliably and let the agronomist worry about interpreting it. made me feel better about my RGB setup tbh