so ive been doing field scouting the old fashioned way for about 6 years now, walking rows, taking notes, all that. my operation is around 800 acres of corn and soy and honestly the manual scouting is just not keeping up anymore, i miss stuff between visits and by the time i catch a nitrogen deficiency or early blight its already spread further than id like.

been looking seriously at the DJI Agras T40 and also the Autel Dragonfish for NDVI monitoring but im not sure if i actually need a dedicated ag drone or if something like a Mavic 3 Multispectral would cover most of my needs without the insane price tag. my main use case right now is monitoring, not spraying, ill probably add spraying capability later once i figure out the workflow.

anyone running multispectral surveys on fields this size? curious how long flights are taking you and what kind of ground sampling distance youre getting at typical survey altitudes. also wondering about the processing pipeline ? are most of you running Pix4Dfields or something else for the maps

800 acres is totally doable with a Mavic 3 Multispectral if monitoring is your only goal right now. i run about 600 acres with mine and a full NDVI survey at 120m altitude takes maybe 2.5 to 3 hours with battery swaps factored in, GSD comes out around 5-6cm which is more than enough to catch stress patterns before they get bad. the reflectance calibration panel thing trips people up at first but once its part of your preflight routine its fine.

Pix4Dfields is what most people around here seem to use and its pretty solid honestly, DroneDeploy also works if youre already in that ecosystem. the M3M is way cheaper to get into than a full ag platform and since you said spraying is a later thing id honestly just start there and upgrade when you need to. dont overthink the first step

yeah this is something i thought about too but i ended up going straight for the T40 because i knew spraying was eventually going to be part of the plan and i didnt want to buy two platforms. in hindsight the T40's mapping mode is honestly just okay, like it works but its clearly not the primary use case for that bird. if i was doing monitor-only i probably wouldve gone multispectral dedicated first.

one thing nobody told me upfront ? the data processing side takes real time to figure out and theres kind of a learning curve on interpreting the NDVI outputs accurately, like what counts as actual stress vs sensor noise or shadow artifacts. first few maps i made i was kind of guessing. might be worth looking into some of the agronomist communities that overlap with drone stuff because they have a different perspective on reading the outputs than the pilot-first crowd