okay so this has been bugging me for a while. running a Mavic 3 Multispectral for field scouting this year and i keep second guessing whether the NDVI data im pulling is actually giving me earlier detection of nitrogen stress or disease pressure versus what i was getting with just good RGB imagery and knowing what to look for visually. like the NDVI maps look impressive and the agronomist loves getting them but when i go back and compare the timeline of when stress shows up in the multispectral data versus when you can just kind of see it starting in a regular RGB pass, the gap isnt always as big as i expected.

maybe im not processing the data right or my ground truthing isnt tight enough. using DJI Terra for the processing and then pulling into QGIS. fields are irrigated pivot corn in Nebraska if that matters. anyone else gone down this rabbit hole or am i just not giving it enough time to see the real difference at scale.

you're not overthinking it, this is a real debate and honestly the answer is kind of both depending on what stress you're chasing. nitrogen stress specifically, RGB can catch it pretty late but NDVI using the red edge band on the M3M starts showing differential uptake issues noticeably earlier in my experience, probably 5-7 days before you'd pick it up visually. disease is different story, some foliar stuff actually shows up in RGB thermal signatures before spectral indices catch it depending on pathogen.

the processing side matters a lot too. are you doing any atmospheric correction or just running raw reflectance? uncorrected NDVI across different flight times can give you noise that makes it look like stress variation when its really just illumination differences between your flight sessions. that might be part of why your comparison is feeling muddy.

i had a similar feeling my first season with multispectral and eventually just talked to the extension agronomist at UNL about it. her take was basically that for most in-season management decisions the timing advantage of multispectral over a trained eye with good RGB is smaller than the vendors imply, but the real value is repeatability and documentation over multiple seasons when you build up enough data to actually see patterns across years. not necessarily helpful for your immediate question but changed how i thought about justifying the cost of the sensor.