so ive been using dronedeploy for about 8 months now doing mostly crop health surveys and some basic elevation models for a local farming co-op. its fine, does what i need, but the subscription cost is starting to bug me especially since i only really use it heavily like 3-4 months out of the year when the growing season is active.

started looking at pix4d and honestly the pricing is kind of all over the place depending on which version you need. pix4dfields looks interesting for the ag stuff but then if i want to do any construction site work on the side id need a different module or just go with pix4dmapper which is way more expensive. dronedeploy just kind of does everything in one package even if it isnt the deepest on any one thing.

anyone actually switched between the two or uses both? mostly curious about the point cloud quality differences if im doing any kind of terrain modeling, and whether the offline processing in pix4d is actually noticeably better than dronedeploys cloud processing. my internet out at some of these farm sites is genuinely terrible so that part does appeal to me.

ran both for a while on a construction monitoring contract and honestly for pure point cloud accuracy pix4dmapper wins but its not like night and day unless youre doing really precise volumetric calculations. for crop stuff i genuinely think dronedeploy is fine and the NDVI outputs are decent enough that the agronomists i work with dont complain.

the offline processing thing is real though. pix4d running locally on a decent workstation is just more reliable when youre in the middle of nowhere. i had dronedeploy uploads fail on me twice because the connection dropped halfway through and had to reshoot parts of a field which was a whole thing. that alone might justify the switch depending on your situation.

i use pix4dfields specifically and its pretty good for what it is but if you want to do anything beyond ag you really do need mapper and yeah the cost stacks up fast. one thing nobody talks about is how long the pix4d processing queue can get if youre on the cloud version, ive had jobs sit for like 2 hours before they even start. local processing solves that but then you need a machine that can handle it which is its own investment