We routinely need to do photogrammetric mapping at very low altitudes in order to obtain sub-cm GSD (for habitat classification and counting/measuring intertidal marine critters and plants). I'm talking as low as 2 m AGL at times, but routinely 5-15 m. For obvious flight safety reasons, many autonomous flight planning and control apps have a minimum altitude that you can set for a photogrammetry mapping flight. In the old Pix4D Capture and in Drone Deploy, it is 10 m, which we can live with for the most part; we use Litchi or hand-fly when we need to go lower.

I'm just planning my first intertidal surveys with a P4R, and of course I must use GS RTK for control. I was amazed to see that the minimum altitude for a 2D mapping mission is 25 m! This seems extreme, and is ironic because with RTK fixed, the P4R's altitude awareness is much more stable and accurate than non-RTK enabled aircraft that must rely on barometric pressure to measure altitude relative to takeoff.

I'd really like to use the P4R for low altitude mapping. I'm hoping there is a way to fly below 25 m AGL.

• First question: does anybody know how to remove the minimum altitude limit?
• If not, I do see where there is a "Relative Height (m)" setting (Default = 0), with an info blurb that says "Set the altitude of the shooting environment relative to the takeoff point. Take off in a safe area". It includes a dynamic indicator of "Actual Flight Altitude" that updates if the setting is changed. As one might intuit, positive values entered in the setting decrease the "Actual Flight Altitude". So presumably if I set "Height (m)" = 25 m (the min) and "Relative Height (m)" = 10 m, even though my takeoff location is within the area to be mapped (so has relative elevation = 0), I should end up with a survey altitude of 15 m AGL (at least that's what the "Actual Flight Altitude" value suggests). Has anybody tried this? My biggest concern is that while this approach may allow me to fly nat the desired altitude, the line and shot spacing will not be adjusted to ensure the proper overlap for 15 m AGL. I suppose I could tweak the overlap settings for 25 m AGL higher to account for this and provide the desired overlap @ my altitude.
• I suppose I could accomplish something similar by doing a terrain aware flight and tweaking the DSM. Rather not go there but will if I have to. In theory this approach would also have the same overlap concerns as method #2. If someone has done this, I'd be interested to hear about it as well.
  

I'm well aware of the safety risks involved with very low altitude flight. Apart from bluffs/cliffs to the inshore of the study area, there is no topographic relief > 2-3 m. Some large boulders, but no trees, poles, powerlines, etc. and no dramatic slope across the area.


Any ideas or constructive advice much appreciated. Please, if you don't have knowledge of the P4R and GS RTK, don't bother posting suggestions that are not relevant.

Thanks!

Bueller?


*crickets*

patiam Posted at 6-13 13:21
Bueller?

sorry , cant help u. Im not used to flying that altitude neither "relative height"

Guille_Meinero Posted at 6-14 04:04
sorry , cant help u. Im not used to flying that altitude neither "relative height"

Thank you @Guille_Meinero. Even an "I dunno" is better than crickets

Hi,

Tested today with UGCS, 5m minimum, not below.

Haven't used UGCS... I don't have the SDK controller, so not sure if I can install it... But thanks for the idea!

I flew some GS RTK test flights yesterday using the relative altitude trick- flight plan altitude set to 25 m minimum, and the relative @ 15, 18, and 20 m to yield 10, 7, and 5 m AGL flights. The flights themselves went perfectly. The results after processing in Pix4D bore out my concerns; even though I cranked the overlap up some (to 90%/85%)  things began to fall apart @ the lower altitudes: @ 10m , 3/185 cameras couldn't be calibrated. @ 7m 63 failed, and at 5m 172.  The line and shot spacing for 25 m AGL just doesn't work @ 5m, even w/ 90%/85% overlap set for 25m.  I'll try even higher overlap next time, maybe even do the math to see what the limits are.

I just ran into this issue today due to the FAA only allowing me to fly at 50 feet.  As an easier solution for my project I will be filing for a waiver to fly at a greater height, but this is a pretty ridiculous feature.  I was told by DJI that is was due to the GS software not being able to process good final results from that height due to image smear etc.  

Jason Chambless Posted at 7-2 06:22
I just ran into this issue today due to the FAA only allowing me to fly at 50 feet.  As an easier solution for my project I will be filing for a waiver to fly at a greater height, but this is a pretty ridiculous feature.  I was told by DJI that is was due to the GS software not being able to process good final results from that height due to image smear etc.

Thanks @Jason Chambless. Looking at my imagery from 15 m AGL I'm beginning to agree that the 25 m minimum may be due to a system limitation.

My 5 mm GSD imagery is smeary in the along-track direction. Not sure if slowing down would help or not, and lighting was not great for these flights. Need to look into this more.

Do you have a contact @ DJI that gave you the info you shared here regarding the limit?

@patiam thanks for your input, I had the exact same issues and considerations. Regarding your tests, it's good you verified that the trick works, achieving at least the desirable flight height. Boosting the overlaps is obviously necessary as, I guess, slowing down the drone proportionally to the flight height reduction for better image quality.
I would be eager to hear more about your progress on the matter, I will definitely share my conclusions asap.

I'll be trying this again in a couple days, flying @ 7 m AGL by setting the 25 m minimum and 18 m relative height. I'll be slowing the speed down to the 1 m/s minimum and bumping overlap up to 90%/90% maximums.

I'll let y'all know how it goes.

7 m AGL intertidal zone flights went well with the settings described in my previous post. Results had ~3 mm GSD, with ~1.5 cm accuracy in XY&Z based on 3 checkpoints surveyed in using RTK GNSS. The blurring/smearing that was a problem in previous very low altitude tests was much reduced by flying @ 1 m/s, and overlap was good (5+ images) for pretty much the whole site by using double-grid flight plan (even though line spacing was driven by the supposed 25 m AGL altitude).

So what does the Relative Height mean? Im heading out tomorrow and have my altitude set at 50m but the the relative height at 0 because I didnt really know what that was for. A quick Youtube search showed a Propeller Aero video on the app and they left it at 0 as well. Do I need to set that as well at 50m??

Slot Tech Posted at 12-1 14:41
So what does the Relative Height mean? Im heading out tomorrow and have my altitude set at 50m but the the relative height at 0 because I didnt really know what that was for. A quick Youtube search showed a Propeller Aero video on the app and they left it at 0 as well. Do I need to set that as well at 50m??

Relative height is an offset to be applied to your TO elevation/altitude to detrmine the P4R's flighht altitude. If your TO location is not the same elevation as your study site, it allows you to achieve the desired altitude AGL in the site.

For example:

• You're taking off from a spot that is 15 m lower than the area to be mapped, and want to map @ 50 m AGL. You enter -15 m in the relative height, and the P4R will fly at the correct altitude in the site (but 65 m AGL from the TO location).
• You're taking off from a place that is 50 m higher than the mapping area, say from a bluff on the coast, but you want to map the beach and sea surface @ 120 m AGL. You put 50 m in RH, and the P4R will fly only 70 m AGL above your TO location and 120 m AGL over the AOI.
• Your TO spot and AOI are the same elevation, but you want to "cheat" the 25 m AGL minimum altitude in GS RTK (which is a silly handicap DJI put there), so you tell the P4R to fly at 25 m and put the difference between that and your desired altitude in RH, which is the topic of the OP.
• A combination of 1 or 2 & 3, you can figure out the math.
• If none of the above apply, set RH = 0.
  

Hope that makes sense. And of course, make sure you're not telling the drone to crash by entering an offset that doesn't agree with the topography!

Hi Patiam,

1. no, I don't think is possible to remove the minimum altitude

2. I use that method in some surveys when I need details, and I tweaked the overlap settings with simple triangles geometry:
if you want to fly at 15m and you need a real 65% overlap you have to set relative height 10m and 79% overlap.
I wrote down a formula to do this:
H height (in this case the minimum is 25m)
RH relative height
OL overlap required
OLS overlap to set

OLS = [(RH / H) * (100 - OL)] + OL

so:  [(10/25)*(100-65)]+65 = 79%

3. Yes, you can tewak the DSM but you have to use the method 2 to have the right overlap.

An advice: don't forget to slow the speed to avoid bad shoots.

Sorry for the poor english. I hope this helps.

Ciao Lowlee-

Thanks for your contribution. Note that the original post was a year and a half ago and most of what you added was discussed and posted already. But thanks for sharing your formula!

Note that since the maximum possible OLS is 90%, there are limits of RH that are determined by OL... For your OL= 65% example, RH maxes out at ~ 18 m. But if OL = 80%, RH can't exceed 12.5 m, and for 85%, 8 m is the max RH.

Oh, sorry...

Yes, you are right!

No worries! Thank you again for sharing your useful formula!

[In practice I usually just max the OL & go as slow as possible (1 m/s) to maximize the chances I'll get what I need... The sites we map at this low altitude are small, so we can afford it.]

patiam Posted at 2022-2-26 22:47
I'll be trying this again in a couple days, flying @ 7 m AGL by setting the 25 m minimum and 18 m relative height. I'll be slowing the speed down to the 1 m/s minimum and bumping overlap up to 90%/90% maximums.

I'll let y'all know how it goes.

Hey, I'm following this thread.  I also messaged you.

ecolish Posted at 4-20 15:38
Hey, I'm following this thread.  I also messaged you.

Don't see a PM. The thread has pretty much run its course. Good info here, but there is probably little useful to add.

PM received & replied.