Good morning,  I just saw that the altitude of the exif is bad. Barometric altitude? Why isn't GPS altitude used? This poses problems for all photogrammetric work.

Thanks

LV_Forestry Posted at 6-7 11:52
[view_image]

My MRK use relative altitude...
My drone was at 395 meters, not 423....

djiuser_lJxE4omQVssJ Posted at 6-7 12:10
Thanks, but where can I do it ?
How choose absolute altitude ?

It's up to you to choose the right datum in your post process. You switch from EGM96 to the one you want, and the deal is done.

GPS altitude is far from accurate, and not reliable.

Read the thread here, the posts from Meta4 explains the different altitudes in EXIF:
https://mavicpilots.com/threads/ ... 37808/#post-1554048

djiuser_lJxE4omQVssJ Posted at 6-7 12:10
My MRK use relative altitude...
My drone was at 395 meters, not 423....

On the screenshot it is the ellipsoidal altitude from the MRK file which is used to make the PPK. Same as before, you start from GRS80 and you go to the datum that interests you

ArcticPhoto Posted at 6-7 12:22
GPS altitude is far from accurate, and not reliable.

Read the thread here, the posts from Meta4 explains the different altitudes in EXIF:

its RTK comrade. not consumer level

LV_Forestry Posted at 6-7 12:23
On the screenshot it is the ellipsoidal altitude from the MRK file which is used to make the PPK. Same as before, you start from GRS80 and you go to the datum that interests you

I don't understand. Can you explain me the method ?

djiuser_lJxE4omQVssJ Posted at 6-7 12:10
My MRK use relative altitude...
My drone was at 395 meters, not 423....

According to the data you present. I see the drone 20m above the ground. 373m HAE and 423m ortho. All is well.

LV_Forestry Posted at 6-7 12:30
According to the data you present. I see the drone 120m above the ground. 423m HAE and 473m ortho. All is well.

Le drone était à 20 mètres au dessus du sol, oui. Mais à une altitude de 385 m, environ...

djiuser_lJxE4omQVssJ Posted at 6-7 12:30
I don't understand. Can you explain me the method ?

Riom, Near your location, the ground is 350m above mean sea level (Ortho)
The geoid RAF09 is +49.4876m at this location. So the HAE of the ground is 400m

MRK says 423m so your drone is at +/- 23m from the ground

In the EXIF there should be written absolute altitude 373m

djiuser_lJxE4omQVssJ Posted at 6-7 12:37
Le drone était à 35 mètres au dessus du sol. Et à une altitude de 385 m, environ...

Could you please share the picture via google drive or something else. it will be simpler to explain

LV_Forestry Posted at 6-7 12:40
Riom, Near your location, the ground is 350m above mean sea level (Ortho)
The geoid RAF09 is +49.4876m at this location. So the HAE of the ground is 400m

Yes, but why Reality Capture dont use absolute altitude ?
How can I read the exif ?

LV_Forestry Posted at 6-7 12:41
Could you please share the picture via google drive or something else. it will be simpler to explain

Which picture ?
The MRK files ?

djiuser_lJxE4omQVssJ Posted at 6-7 12:43
Yes, but why Reality Capture dont use absolute altitude ?
How can I read the exif ?

Honestly I have zero clues from Reality Capture. If you send me the photo, I will be able to explain to you by opening the EXIF with EXIF tool

djiuser_lJxE4omQVssJ Posted at 6-7 12:45
Which picture ?
The MRK files ?

The JPG from the drone SD card

LV_Forestry Posted at 6-7 12:46
The JPG from the drone SD card

Picture

djiuser_lJxE4omQVssJ Posted at 6-7 12:49
Picture

You must grant access to the drive. I sent you PM

Picture public

So, here is the EXIF :




When you started the drone, its relative altitude went to zero.
When the photo was taken you were 25 m from this starting point along the Z axis.

The indicated altitude is 423.3 above sea level. So the ground is at 398.3 above sea level. And it is wrong. It is 50m lower.

I am trying to find out why. Because according to the DJI doc it should be the ortho altitude that should be displayed.

It doesn't change much for you. You start from WGS84 GRS80, then you transform to RGF93 RAF09 to have a correct raster.

LV_Forestry Posted at 6-7 13:27
So, here is the EXIF :
[view_image]

Well, I understand...
But I don't kow why, with the RTK, the altitude is not directly corrected.

I have to change th output coordinates système in RGF93. But I can't see RAF09.

OK, I see. It seems to be a bug. I have a dataset without RTK and it is the AMSL altitude (ortho) that is displayed. It seems that when you put the RTK module, it goes into HAE. But in the EXIF it retains the designation Above sea level.

You can do the test, do a flight mission with the RTK module connected to the centipede network. Redo the same mission by removing the RTK module. If we follow this reasoning you should have two offset dataset of the geoid value at this place.

There is indeed a moderator who will find an answer for us. Otherwise you send an email to enterprise@dji.com, remembering to share the answer with us.

djiuser_lJxE4omQVssJ Posted at 6-7 13:32
Well, I understand...
But I don't kow why, with the RTK, the altitude is not directly corrected.

Yes sorry go to 9538 or 5698 IGN69.



IGN78 is for Corsica.

LV_Forestry Posted at 6-7 13:41
Yes sorry go to 9538 or 5698 IGN69.

I don't change anything to change the coordinate system ouput.

A ground point is above 400 meter...

djiuser_lJxE4omQVssJ Posted at 6-7 13:55
I don't change anything to change the coordinate system ouput.

A ground point is above 400 meter...

Then you need to implement the geoid somewhere.
https://s3-eu-west-1.amazonaws.c ... tg/fr_ign_RAF09.tif

LV_Forestry Posted at 6-7 13:57
Then you need to implement the geoid somewhere.
https://s3-eu-west-1.amazonaws.com/download.agisoft.com/gtg/fr_ign_RAF09.tif

RealityCapture does not support geoid model.
Can I convert my heights into a system which uses ellipsoidal model.
Reedtoolbox for example ?

djiuser_lJxE4omQVssJ Posted at 6-7 14:05
RealityCapture does not support geoid model.
Can I convert my heights into a system which uses ellipsoidal model.
Reedtoolbox for example ?

Do not waste your money in Redtoolbox. Produce your raster in HAE, then come back to me, i will explain how to get the correct elevation with free software (QGIS)

And think about serious survey mapping software as Metashape.

LV_Forestry Posted at 6-7 14:11
Do not waste your money in Redtoolbox. Produce your raster in HAE, then come back to me, i will explain how to get the correct elevation with free software (QGIS)

And think about serious survey mapping software as Metashape.

Is right ?

djiuser_lJxE4omQVssJ Posted at 6-7 14:22
Is right ?

no

These are the estimates of the precision on X Y and Z
+/- 1.5cm on X
+/- 1.2cm on Y
+/- 2.1cm on Z

What you just did reminds me of this:

423.78 - 49.4876 = 374.2924 m
HAE - geoid = Ortho (AMSL)

LV_Forestry Posted at 6-7 14:28
no

These are the estimates of the precision on X Y and Z

LV_Forestry Posted at 6-7 14:28
no

These are the estimates of the precision on X Y and Z

Vous êtes français ?
Si oui, n'existe pas une ligne de commande qui me permet de corriger mon fichier MRK ?

Is there a command line that allows me to fix my MRK file? And add X meters to my altitude.

LV_Forestry Posted at 6-7 14:34
423.78 - 49.4876 = 374.2924 m
HAE - geoid = Ortho (AMSL)

Ou trouvez cela ?

LV_Forestry Posted at 6-7 14:11
Do not waste your money in Redtoolbox. Produce your raster in HAE, then come back to me, i will explain how to get the correct elevation with free software (QGIS)

And think about serious survey mapping software as Metashape.

Mateshape is awful to clean the points artefacts...

djiuser_lJxE4omQVssJ Posted at 6-7 14:35
Vous êtes français ?
Si oui, n'existe pas une ligne de commande qui me permet de corriger mon fichier MRK ?

You can write a Python script for that, then input the offset value as argument. But this is DIY, not surveying.
The altitude i gave you came from the geoid RAF09 (IGN69).

In Metashape it works very well :

1/


2/


3/


voila !

LV_Forestry Posted at 6-7 13:27
So, here is the EXIF :
[view_image]

Hello LV_Forestry,

the above mean sea level is misleading, the altitude stored is above ellipsoid..The MRK file clearly stores HAE and not mean sea level....

PaulKofi Posted at 6-7 14:44
Hello LV_Forestry,

the above mean sea level is misleading, the altitude stored is above ellipsoid..The MRK file clearly stores HAE and not mean sea level....

Hi Paul,

Yes i notice that. Another victory for DJI.

djiuser_lJxE4omQVssJ Posted at 6-7 14:35
Vous êtes français ?
Si oui, n'existe pas une ligne de commande qui me permet de corriger mon fichier MRK ?

Bonjour,

je peux t'aider. Le fichier MRK contient les coordonnées precises RTK de chaque photo dans le système de coordonnées WGS84 avec elevation ellipsoïdale. Si tu veux des elevations au niveau de la mer (geoide) tu dois appliquer un modele de geoide pour corriger celles-ci. Le modele Français est le RAF09 ou RAF18 (plus récent). Si tu partages le fichier MRK, je peux te les transformer.....

PaulKofi Posted at 6-7 14:54
Bonjour,

je peux t'aider. Le fichier MRK contient les coordonnées precises RTK de chaque photo dans le système de coordonnées WGS84 avec elevation ellipsoïdale. Si tu veux des elevations au niveau de la mer (geoide) tu dois appliquer un modele de geoide pour corriger celles-ci. Le modele Français est le RAF09 ou RAF18 (plus récent). Si tu partages le fichier MRK, je peux te les transformer.....

You will be surprised as i am but there is RAF20 now

https://geodesie.ign.fr/contenu/ ... metropole/RAF20.tac

LV_Forestry Posted at 6-7 14:56
You will be surprised as i am but there is RAF20 now

https://geodesie.ign.fr/contenu/fichiers/documentation/grilles/metropole/RAF20.tac

Incredibile! Must not be much different from RAF18 but things change...
The average difference between RAF20 and RAF18 is 4 mm, max diff 18 cm, min diff -22 cm and stdev 1.9 cm