This has been raised before, see: https://forum.dji.com/thread-117047-1-1.html

My home point was 130m above sea level (Ordanance Survey data). I launched and climbed to a barometric altitude of 54.8m (i.e. 184.8m above sea level). I flew to a point where the ground below was 55m above sea level. I then took a jpeg photo. Hence the actual altitude of the Phanton 4 when the photo was taken was 184.8m above sea level. On examining the EXIF it contains:
        <drone-dji:AbsoluteAltitude>+58.71</drone-dji:AbsoluteAltitude>
         <drone-dji:RelativeAltitude>+54.90</drone-dji:RelativeAltitude>

         <exif:GPSVersionID>0.0.0.0</exif:GPSVersionID>
         <exif:GPSLatitude>50,58.5364N</exif:GPSLatitude>
         <exif:GPSLongitude>1,7.8516W</exif:GPSLongitude>
         <exif:GPSAltitudeRef>0</exif:GPSAltitudeRef>
         <exif:GPSAltitude>58707/1000</exif:GPSAltitude>


It seems likely that the RelativeAltitude 54.9m is the barometric alttitude (i.e. the altitude above the home point). The GPSAltitude of 58.71 above sea level is completely wrong: GPS could not be out by 126m surely.

It seems more likely that what the drone stores in GPSAltitude is actually the GPS Altitude minus GPS altitude of the home point. That would mean that the discrepancy between barometric and GPS altitudes was only 4m or so.

What does DJi say?

What does DJi say?
DJI hardly ever say anything here.

The GPSAltitude of 58.71 above sea level is completely wrong: GPS could not be out by 126m surely.
126 metres is a bit extreme but it is possible.  Errors of +/- 200 feet are common.

It seems more likely that what the drone stores in GPSAltitude is actually the GPS Altitude minus GPS altitude of the home point. That would mean that the discrepancy between barometric and GPS altitudes was only 4m or so.
GPS altitude is the altitude supplied by GPS without adding or subtracting anything.
GPS is terribly inaccurate for altitude which is why the drone doesn't use GPS for altitude data.
GPS altitude swings a lot and can be completely different just an hour later.

Don't get too excited about the data from just one photo.
Check data from several photos taken on a number of days.
If you take a test shot before launching, that saves the complication of arithmetic.

Labroides Posted at 2018-6-25 05:19
What does DJi say?
DJI hardly ever say anything here.

GPS altitude is the altitude supplied by GPS without adding or subtracting anything - Well, how do you know? I am asking the question because I want to hear from DJI.

I checked another jpeg, which I took at home on another day before the drone had taken off. It has RelativeAltitude 0.0, GPSAltitude 3.0, actual altitude above sea level 50m. I would accept a GPS error of 50m in altitude, but not one of 120m. Hence I think my suggestion is more likely to be true. But I can check this by going to a realy high place and taking a photo on the ground.

pedro05 Posted at 2018-6-25 05:46
GPS altitude is the altitude supplied by GPS without adding or subtracting anything - Well, how do you know? I am asking the question because I want to hear from DJI.

I checked another jpeg, which I took at home on another day before the drone had taken off. It has RelativeAltitude 0.0, GPSAltitude 3.0, actual altitude above sea level 50m. I would accept a GPS error of 50m in altitude, but not one of 120m. Hence I think my suggestion is more likely to be true. But I can check this by going to a realy high place and taking a photo on the ground.

Well, how do you know? I am asking the question because I want to hear from DJI.
Good luck waiting for DJI to answer, it's unlikely to happen.
The moderators here don't have the knowledge to answer and no-one else from DJI will.
Even if they did, the answer is as likely to be wrong as correct

How do I know???
I've had over 20 years using GPS professionally, and I'm a professional drone photographer.
I answer hundreds of enquiries here to help users like you but if you you think you know better ......

I've given you accurate information and a way to check it.
Check photos taken on the same point on a number of different days and you'll find the altitudes are all over the place.
You can't work out what's behind it from only one or two.
Whether you believe me makes no difference to me.

And if you're still skeptical and want something from another source, here's what Garmin have to say about GPS altitude:
Here's what Garmin have to say about GPS altitude accuracy and why GPS altitude data is useless.
How accurate is the GPS elevation reading?
GPS heights are based on an ellipsoid (a mathematical representation of the earth's shape), while USGS map elevations are based on a vertical datum tied to the geoid (or what is commonly called mean sea level). Basically, these are two different systems, although they have a relationship that has been modeled.

The main source of error has to do with the arrangement of the satellite configurations during fix determinations. The earth blocks out satellites needed to get a good quality vertical measurement. Once the vertical datum is taken into account, the accuracy permitted by geometry considerations remains less than that of horizontal positions. It is not uncommon for satellite heights to be off from map elevations by +/- 400 ft. Use these values with caution when navigating.

Labroides Posted at 2018-6-25 05:57
Well, how do you know? I am asking the question because I want to hear from DJI.
Good luck waiting for DJI to answer, it's unlikely to happen.
The moderators here don't have the knowledge to answer and no-one else from DJI will.

Now you're talking! Thanks, I had forgotten about geoids. In the UK it is the difference between OSGB36 and WGS84, and in my neck of the woods you have to subtract 46 metres from a WGS84 height to convert it to OSGB36.  Sadly, that makes the observed error worse. It is now 184.8 - (58.7 - 46) = 172.1m (or 550 ft). Wish that correction were the other way around, but it isn't.

I did another test today, putting my Satmap Active 10 hand-held GPS on the ground next to the Phantom 4 at an altitude of 192m above sea level and taking a photo after the home point had been set but before take off. Satmap got the altitude right to within 2m, whereas drone GPSAltitude was 104m low.

I also tried flying below home point altitude: GPSAltitude remained positive, so you are right in saying my speculation that this EXIF value might be relative to the home point was incorrect.

pedro05 Posted at 2018-6-25 09:57
Now you're talking! Thanks, I had forgotten about geoids. In the UK it is the difference between OSGB36 and WGS84, and in my neck of the woods you have to subtract 46 metres from a WGS84 height to convert it to OSGB36.  Sadly, that makes the observed error worse. It is now 184.8 - (58.7 - 46) = 172.1m (or 550 ft). Wish that correction were the other way around, but it isn't.

I did another test today, putting my Satmap Active 10 hand-held GPS on the ground next to the Phantom 4 at an altitude of 192m above sea level and taking a photo after the home point had been set but before take off. Satmap got the altitude right to within 2m, whereas drone GPSAltitude was 104m low.

I did another test today, putting my Satmap Active 10 hand-held GPS on the ground next to the Phantom 4 at an altitude of 192m above sea level and taking a photo after the home point had been set but before take off. Satmap got the altitude right to within 2m, whereas drone GPSAltitude was 104m low.

One factor to consider when using the Satmap 10 is that it gets altitude data two ways.
From the Satmap Active 10 User guide  (p31)
http://cdn.satmap.com.s3-eu-west-1.amazonaws.com/FULL-150-UG-EN-1105_WEB.pdf

Elevation Data
Altitude from maps/Altitude from GPS
• Elevation data is found on all map cards (except Ordnance Survey where the elevation data is taken from the base mapping of the units).
• GPS elevation data is also available and is accurate to ± 50m

If you got altitude correct within 2 metres, you probably were getting altitude from the map data card.
Select Altitude Data from GPS to see how variable GPS altitude really is.

In summary, GPS is really dodgy when it comes to altitude.

Labroides Posted at 2018-6-25 14:24
I did another test today, putting my Satmap Active 10 hand-held GPS on the ground next to the Phantom 4 at an altitude of 192m above sea level and taking a photo after the home point had been set but before take off. Satmap got the altitude right to within 2m, whereas drone GPSAltitude was 104m low.

One factor to consider when using the Satmap 10 is that it gets altitude data two ways.

Thanks. I have that manual in the box for the Active 10 but it is printed so small I can barely read it, so good to see it big on my desktop PC. I checked my Active 10 settings and it is set for "Altitude from GPS". It has an Ordnance Survey map card installed which, the manual says, does not have elevation data.

It isn't clear from manual if the Active 10 transforms WGS84 heights to OSGB36 ones when set to display position coordinates in OSGB. I've asked Satmap.

Welcome to the club of confused users, i tried myself to understand in this thread.
If you manage to figure out why the Phantom is so far off compared to other equipment please try to make me understand with other words as i am still confused of the other replies i got so far.

KedDK Posted at 2018-6-25 22:59
Welcome to the club of confused users, i tried myself to understand in this thread.
If you manage to figure out why the Phantom is so far off compared to other equipment please try to make me understand with other words as i am still confused of the other replies i got so far.

The Phantom matches the variability I see in altitude measured by my Garmin handhelds (those without a barometer).
Mine has showed me being >200 ft below sea level while walking on the beach.
Many modern GPS units incorporate a barometer to be able to show useful altitude data - but not useful in a pressurised aircraft.

Hi,

Please join us on ct2gnh@hotmail.com ! Send us a email with your email, so we can keep you up with this problem !
We are already 10 persons asking for a soluction to dji, and we need more people ! United we will win !

This last firmware upgrade didnt repair anything ! GPS Altitude on Fotos is not ok !!

Join us !

Labroides Posted at 2018-6-26 00:03
The Phantom matches the variability I see in altitude measured by my Garmin handhelds (those without a barometer).
Mine has showed me being >200 ft below sea level while walking on the beach.
Many modern GPS units incorporate a barometer to be able to show useful altitude data - but not useful in a pressurised aircraft.

Correct me if I am wrong but there is no way to correct the GPS altitude usign the barometric altitude.?

And if i am not mistaken, a simple fix to the problem is to simple overwrite the GPS altitude in the exif file, by adding up the barometric hight and the approximate altitude of the area one was flying in. Of course there is again an error included but i would guess its way smaller than the one from inaccurate GPS signal.  

mKratos Posted at 11-19 03:27
Correct me if I am wrong but there is no way to correct the GPS altitude usign the barometric altitude.?

And if i am not mistaken, a simple fix to the problem is to simple overwrite the GPS altitude in the exif file, by adding up the barometric hight and the approximate altitude of the area one was flying in. Of course there is again an error included but i would guess its way smaller than the one from inaccurate GPS signal.

Although DJI call it GPS altitude in the Exif info, it does not come from GPS at all.
More recent testing has found that it comes from the barometer and is inaccurate because air pressure varies and DJI's formula assumes a standard air pressure.