Incorrect EXIF GPS and Elevation

[Paulo Lobo]

exiftool

Can we make a python aplication to auto copy the relative altitude to gps altitude in the fotos ??

 

[sar104]

So the relative altitude is 31.60meters, this is the real one that can be use ?

The relative altitude (barometric) is going to be pretty accurate (to within a couple of meters) but you will have to add your take off elevation above MSL to get altitude above MSL.

 

[Paulo Lobo]

The relative altitude (barometric) is going to be pretty accurate (to within a couple of meters) but you will have to add your take off elevation above MSL to get altitude above MSL.

Thats is to much to do, so this is what dji must repair... I dont understand is why the fotos from yesterday are ok, and today are not... its all about positive and negative values, so if we calibrate before we fly can it work out ?

 

[Keule]

Can we make a python aplication to auto copy the relative altitude to gps altitude in the fotos ??

Read this article: Sorting out DJI's bad GPS Altitude tagging

 

[Paulo Lobo]

I read that, but it give lot of work for 800 fotos...
Please check this foto from yesterday !


Look at gps altitude and relative altitude please...

DJI_0001.JPG

 

[sar104]

Thats is to much to do, so this is what dji must repair... I dont understand is why the fotos from yesterday are ok, and today are not... its all about positive and negative values, so if we calibrate before we fly can it work out ?

I read that, but it give lot of work for 800 fotos...
Please check this foto from yesterday !


Look at gps altitude and relative altitude please...

DJI_0001.JPG

GPS Altitude Ref : Above Sea Level​

Absolute Altitude : +37.91​

Relative Altitude : +40.80​

GPS Altitude : 37.9 m Above Sea Level​

What about this is so hard to understand? In this file the absolute barometric altitude is positive, and so it is also reported in the "GPS Altitude" field.

 

[sar104]

Read this article: Sorting out DJI's bad GPS Altitude tagging

Yes - but that article is also wrong in stating that the "GPS Altitude" is actually GPS data, rather than barometric data.

 

[Paulo Lobo]

GPS Altitude Ref : Above Sea Level​

Absolute Altitude : +37.91​

Relative Altitude : +40.80​

GPS Altitude : 37.9 m Above Sea Level​

What about this is so hard to understand? In this file the absolute barometric altitude is positive, and so it is also reported in the "GPS Altitude" field.

Yes ! But how did the mavic 2 pro yesterday did this good job and today is not doing this good job and selecting the AltitudeRef to below sea level ?
This is a barometer problem ? Or its the firmware doing that ?

 

[sar104]

Yes ! But how did the mavic 2 pro yesterday did this good job and today is not doing this good job and selecting the AltitudeRef to below sea level ?
This is a barometer problem ? Or its the firmware doing that ?

No - it's just changes in atmospheric pressure. The barometer measures the pressure and then uses a standard model of the atmosphere (pressure as a function of altitude) to estimate altitude. But as the weather changes, so does the atmospheric pressure and so it is not going to be consistently correct. Low pressure leads to high altitude estimates and high pressure gives low estimates. Since you are near sea level, high atmospheric pressure can actually lead to negative altitude values, as you saw in the earlier image file. The last one that you posted was obviously taken when the atmospheric pressure was lower.

 

[Paulo Lobo]

Dji is working on this since this or even before....

Its so bad...

 

[sar104]

Dji is working on this since this or even before....

Its so bad...


View attachment 60717

The actual GPS altitude data, which are accurate to within a few meters, are in the DAT flight log. Anyone serious about this kind of work should have no problem pulling them from there, even if it is not the ideal situation.

 

[Paulo Lobo]

I need to check how the log is.
Remember, a mapping 3d of a tower can take 800...900...1200 fotos, in many altitudes...so i can not do that foto by foto...

 

[sar104]

I need to check how the log is.
Remember, a mapping 3d of a tower can take 800...900...1200 fotos, in many altitudes...so i can not do that foto by foto...

Luckily you can automate that kind of thing.

 

[Paulo Lobo]

Luckily with the emails and reports im sending to DJI in many sites, they will stop everything and solve this problem very fast...
Or else we return back or drones !

Its mavic 2 pro, mavic 2 zoom, and Mavic enterprise with this problem !

Peolpe are not happy...

 

[sar104]

Luckily with the emails and reports im sending to DJI in many sites, they will stop everything and solve this problem very fast...
Or else we return back or drones !

Its mavic 2 pro, mavic 2 zoom, and Mavic enterprise with this problem !

Peolpe are not happy...

The Mavic Pro and Phantoms do this exactly the same way last time I looked.

 

[Paulo Lobo]

The Mavic Pro and Phantoms do this exactly the same way last time I looked.

No they dont !
I make mapping with mavic 1 pro and phantom 4 pro and there is no problem !!

Look here, fotos from mavic pro 1 !


DJI_0036.JPG and 1 more file

 

[sar104]

No they dont !
I make mapping with mavic 1 pro and phantom 4 pro and there is no problem !!

Look here, fotos from mavic pro 1 !


DJI_0036.JPG and 1 more file

And they have exactly the same fields:

File Name : DJI_0036.JPG​

GPS Altitude Ref : Above Sea Level​

Absolute Altitude : +20.55​

Relative Altitude : +26.50​

GPS Altitude : 20.5 m Above Sea Level​

File Name : DJI_0046.JPG​

GPS Altitude Ref : Above Sea Level​

Absolute Altitude : +48.45​

Relative Altitude : +54.40​

GPS Altitude : 48.4 m Above Sea Level​

Absolute altitude and GPS altitude are reporting the same value, which is the barometrically derived absolute altitude.

 

[Meta4]

if we calibrate before we fly can it work out ?

There's nothing you could calibrate that would make any difference to altitude data.

 

[TheAllrounder]

"This is so so so bad for dji image..."

To maybe 10 people worldwide.

I really have to agree with Paulo here. Just look at the current "Enterprise" market line which exactly targets people who want to do surveying with UAVs. However, the question still remains: Why is there an error? I understood that the EXIF file shows more than e.g. Agisoft can read. Nevertheless, I cannot believe that the problem is a calculation error. In my case, I was using the Mavic to survey crop fields. Now the Altitude in my images differs by about 55m to my GPS measurements on the ground. When I add the relative altitude to the GPS altitude as @Kildir outlined, I am nowhere close to the actual altitude. Can someone explain what is causing the error?

 

[sar104]

I really have to agree with Paulo here. Just look at the current "Enterprise" market line which exactly targets people who want to do surveying with UAVs. However, the question still remains: Why is there an error? I understood that the EXIF file shows more than e.g. Agisoft can read. Nevertheless, I cannot believe that the problem is a calculation error. In my case, I was using the Mavic to survey crop fields. Now the Altitude in my images differs by about 55m to my GPS measurements on the ground. When I add the relative altitude to the GPS altitude as @Kildir outlined, I am nowhere close to the actual altitude. Can someone explain what is causing the error?

You are going to have to be much more precise in that question. Are you saying that you are adding the relative altitude from the EXIF file to the absolute altitude of the takeoff point, or something else? What do you mean by "actual altitude"?

Please read post #16.