Constant Height over long distance?

Joined
Jun 23, 2022
Messages
10
Reaction score
0
Age
33
Hey there,

when I flew my phantom 4 over a distance of several kilometers last weekend, I was wondering weather DJI Phantom 4 with DJI Go 4 App sticks to a certain height above ground level.

In theory, after every 1 km you have 8 cm of drop due to earth's curvature. Taking this into account, would the drone fly 800 cm higher when it could fly 100 km?
Or would it stick to a fixed altitude above ground level referenced to the globe, or the atmospheric pressure at start?

So, the final question would be whether it is theoretically possible to map the entire earth as a globe, or whether photogrammetry software is more likely to give you a flat earth ? :tearsofjoy:


Cheers
SvSz
 
Hey there,

when I flew my phantom 4 over a distance of several kilometers last weekend, I was wondering weather DJI Phantom 4 with DJI Go 4 App sticks to a certain height above ground level.

In theory, after every 1 km you have 8 cm of drop due to earth's curvature. Taking this into account, would the drone fly 800 cm higher when it could fly 100 km?
Or would it stick to a fixed altitude above ground level referenced to the globe, or the atmospheric pressure at start?

So, the final question would be whether it is theoretically possible to map the entire earth as a globe, or whether photogrammetry software is more likely to give you a flat earth ? :tearsofjoy:


Cheers
SvSz

It establishes and maintains altitude based on its internal barometer.
 
  • Like
Reactions: Harleydude
Hey there,

when I flew my phantom 4 over a distance of several kilometers last weekend, I was wondering weather DJI Phantom 4 with DJI Go 4 App sticks to a certain height above ground level.
Your Phantom has no way to tell how high it is above whatever is below it.
It only knows its height relative to the launch point.
 
Hey there,

when I flew my phantom 4 over a distance of several kilometers last weekend, I was wondering weather DJI Phantom 4 with DJI Go 4 App sticks to a certain height above ground level.

In theory, after every 1 km you have 8 cm of drop due to earth's curvature. Taking this into account, would the drone fly 800 cm higher when it could fly 100 km?
Or would it stick to a fixed altitude above ground level referenced to the globe, or the atmospheric pressure at start?

So, the final question would be whether it is theoretically possible to map the entire earth as a globe, or whether photogrammetry software is more likely to give you a flat earth ? :tearsofjoy:


Cheers
SvSz

Re the earth's curve: You're out by a factor of 100... it's actually 785 metres per 100 km. As barometric pressure follows the curvature of the earth, so too would your drone. That said, the home-point elevation my Phantom returns to is rarely zero - and sometimes more than 2 metres out, so who knows what it would do over 100 kms?
 
It establishes and maintains altitude based on its internal barometer.
So they fly among isobaric lines, no matter how they are distributed AGL. I expected that. Is there a mode to disable this?

Therefore, I assume the result would be a flat earth actually, when the photogrammetry does not take this problem into account.

Re the earth's curve: You're out by a factor of 100... it's actually 785 metres per 100 km. As barometric pressure follows the curvature of the earth, so too would your drone. That said, the home-point elevation my Phantom returns to is rarely zero - and sometimes more than 2 metres out, so who knows what it would do over 100 kms?

You're right, I assumed this ratio to be linear, in fact it is ofc more exponential. I think they do quiet a good job at HTTPS-LINK: earthcurvature.com. However, not sure about the results after 1000 km.

Thanks for all the replies so far!!
 
Last edited:
You cannot disable the barometer unless you maybe know what wire to disconnect -- if it will even take off at all at that point. And, as Meta4 explained, the barometer resets to zero (or thereabouts) at startup, so altitude is always relative to your take-off point. If you fly along autonomously maintaining 200 ft altitude and approach a 300 ft hill, you will crash into that hill. Basically, disconnecting the barometer would not help and doing so would make no sense.
 
Last time I was reading something about this equation, I think I remember approximately 8 inches per mile for curvature of earth. Is my memory off?
Yes, I see the 8 inches per mile figure quoted on the internet. That's 12.4 cm per km or 12.4 metres per 100 km... very different from 785 metres, which seems like a lot! Yet, two different 'earth curvature' calculators I found online calculate the 785 metre figure. I don't know what's correct.
 
While not directly an answer to the question I’ll throw this in…

If you program a mission with Litchi, you can specify waypoints to be a certain altitude above the terrain. Of course it’s not based on any kind of sensor input from the drone. Just terrain data from the maps they use in the mission planning software.

FWIW. :)
 
You cannot disable the barometer unless you maybe know what wire to disconnect -- if it will even take off at all at that point. And, as Meta4 explained, the barometer resets to zero (or thereabouts) at startup, so altitude is always relative to your take-off point.

Okay i would not go so far too mess with the hardware, yet, since the P4 is still the best drone I have right now... Hopefully, this is changes until next year. getting off topic now, haha. ;)

Does anyone know why the drone does not use altitude information from GPS? I mean, they are always in the footage's meta data so... the drone actually knows how high above sea level it is...

If you fly along autonomously maintaining 200 ft altitude and approach a 300 ft hill, you will crash into that hill. Basically, disconnecting the barometer would not help and doing so would make no sense.

Sure thing, my flight was located in a near flat terrain with almost no differences in heights / almost no profile. Maybe 6 - 10 feet difference over 2 km.


Pygar70

Yes, I see the 8 inches per mile figure quoted on the internet. That's 12.4 cm per km or 12.4 metres per 100 km... very different from 785 metres, which seems like a lot! Yet, two different 'earth curvature' calculators I found online calculate the 785 metre figure. I don't know what's correct.

If you want to calculate the curvature drop locally, say 1 km from your position, you can always take 8 cm (CENTIMETERS!) after 1 km. That's the ratio to go. But don't stress this out too much, since it's not actually accurate. It's getting more inaccurate when you use farther distances like 100 km! Even at 2 km you will not get 16 cm but rater 31 cm already. This is, because 8 cm / 1 km is a clear linear relationship, whereas the curvature of the earth is not linearly dropping. It sounds complicated but it's actually quiet trivial. The curvature of a ball/sphere is always exponentially dropping, otherwise it wouldn't be a ball. The people from Earthcurvature make a good job, yet they also don't have the best equation for very long distances (see note at the bottom of the page).

bsartist

While not directly an answer to the question I’ll throw this in…

If you program a mission with Litchi, you can specify waypoints to be a certain altitude above the terrain. Of course it’s not based on any kind of sensor input from the drone. Just terrain data from the maps they use in the mission planning software.

Thanks for that! I guess they use google earth elevation data. But you can also upload DEMs into your Litchi mission, pretty nice software to plan flights. Love it!

__________________________________________________________________________________________


So, when we now agree the drone does not fly straight, but can rather change its height above ground level (AGL) along a longer course, would you say my experiment is possible?
I repeat my experiment:

- I wanted to find out if I can see Earth's curvature in a photogrammetry 3D model (with the extent of several km) after loading in missions from my flight.
- Therefore I wanted to compare Point A at 80.00 m above sea level with Point B also at 80.00 above sea level (ASL).
- Point A und Point B are separate by 1 to 3 km. This would result in ...

1 km = 0,08 m
2 km = 0,31 m
3 km = 0,71 m

... drop in Elevation from Point A to Point B caused by Earth Curvature, although they are both at 80.00 m ASL.


Cheers
SvSz
 
Last edited:
Does anyone know why the drone does not use altitude information from GPS?
Because air pressure gives better precision and is convenient.
In real planes the altimeter also uses a barometric sensor.
I mean, they are always in the footage's meta data so... the drone actually knows how high above sea level it is...
Not really. - Here are some examples of the "height above sea level" shown in image metadata:
i-wMWSXK7-X2.jpg


Although DJI label the metadata field "GPS Altitude" it does not come from GPS and is completely useless.
It's derived from air pressure using a flawed algorithm that doesn't account for the normal fluctuations in air pressure.
Easy to test for yourself with a few shots taken at the same height on different days.

- I wanted to find out if I can see Earth's curvature in a photogrammetry 3D model (with the extent of several km) after loading in missions from my flight.
- Therefore I wanted to compare Point A at 80.00 m above sea level with Point B also at 80.00 above sea level (ASL).
- Point A und Point B are separate by 1 to 3 km. This would result in ...

1 km = 0,08 m
2 km = 0,31 m
3 km = 0,71 m
In most places on earth, the terrain isn't going to be flat enough to show such small elevation differences over that distance.
 
  • Like
Reactions: Pygar70
Because air pressure gives better precision and is convenient.
In real planes the altimeter also uses a barometric sensor.

Not really. - Here are some examples of the "height above sea level" shown in image metadata:


Although DJI label the metadata field "GPS Altitude" it does not come from GPS and is completely useless.
It's derived from air pressure using a flawed algorithm that doesn't account for the normal fluctuations in air pressure.
Easy to test for yourself with a few shots taken at the same height on different days.

Very useful information! Grazie mille! This actually fits with my observations. This seems to be a key problem why you cannot use images from different days in photogrammetry software.

In most places on earth, the terrain isn't going to be flat enough to show such small elevation differences over that distance.

I believe it is not important for the surface to be flat at all. All that is important is your accuracy of ASL data for Point A and B. What happens in between is practically insignificant. But sure, It would be nice to have a distance of 5 or more km flown, to get a greater dip.
 
This seems to be a key problem why you cannot use images from different days in photogrammetry software.
What software would you have the problem with?
I never had a problem with Metashape.
It ignores DJI's GPS altitude and just uses the relative altitude.
If you need to match elevations with your local datum, that's solved by using a few ground control points.
 
What software would you have the problem with?
I never had a problem with Metashape.
It ignores DJI's GPS altitude and just uses the relative altitude.
If you need to match elevations with your local datum, that's solved by using a few ground control points.

Exactly with metashape I had this problem. Different days or different drones. Both seems problematic independently. But this is getting off topic I think.
 
Hey there,

when I flew my phantom 4 over a distance of several kilometers last weekend, I was wondering weather DJI Phantom 4 with DJI Go 4 App sticks to a certain height above ground level.

In theory, after every 1 km you have 8 cm of drop due to earth's curvature. Taking this into account, would the drone fly 800 cm higher when it could fly 100 km?
Or would it stick to a fixed altitude above ground level referenced to the globe, or the atmospheric pressure at start?

So, the final question would be whether it is theoretically possible to map the entire earth as a globe, or whether photogrammetry software is more likely to give you a flat earth ? :tearsofjoy:


Cheers
SvSz
As already purported in this thread, the P4 would follow the curve of the Earth because it maintains its altitude based on barometer data, which is relatively consistent within a certain area. Unfortunately, the curvature of the Earth would be the least of your problems. Natural changes in elevation - like hills and mountains - would most certainly, eventually crash your drone.

Two terms you should become familiar with; "AGL" and "Terrain Awareness." DJI products are NOT Terrain Aware. They will happily maintain 400' AGL (Above Ground Level) and crash into the side of a hill or mountain without so much as slowing down. If your drone is moving fast enough, the Object Avoidance will be of little help. The solution is to purchase third-party software that IS terrain aware. Most of these programs (called "apps" by kids these days) utilize SRTM (Shuttle RADAR Topology Mission) data to program terrain data into a given waypoint flight/mission. This will allow your drone to follow the natural curves in the terrain, but of course will do nothing should you encounter a tall building or tower.

D
 

Recent Posts

Members online

No members online now.

Forum statistics

Threads
143,066
Messages
1,467,357
Members
104,935
Latest member
Pauos31