Labroides Posted at 9-19 06:58
I suspect you have it all wrong.
Do you want to explain your issue?
Why would you say that you can't do photogrammetry with a Mavic 2 when so many other users do exactly that?

Photogrammetry is a long procedure.  DJI Mavic 2 PRO shots got it right initially on the first stage of pix4D processing but after the second last firmware update, i am only getting 65% -75% calibrated cameras no matter what i do !!  The side and front overlap were placed > 75% and heights were kept at 75 meters -85 meters.  A 3000 $ PiXD software with a 1700 $  DJI drone must do this type of post processing  jobs without much tuning and fuss  !!
However i didn't try the last firmware for the fear of what i experienced with my DJI SPARK where i lost my motherboard after the firmware update !! )

can somebody explain, what is happening with latest firmware update, photogrammetry wise? does m2p images/exif works ok or not in regular used software (agisoft, pix4d etc)?

S.J Posted at 9-19 09:53
Photogrammetry is a long procedure.  DJI Mavic 2 PRO shots got it right initially on the first stage of pix4D processing but after the second last firmware update, i am only getting 65% -75% calibrated cameras no matter what i do !!  The side and front overlap were placed > 75% and heights were kept at 75 meters -85 meters.  A 3000 $ PiXD software with a 1700 $  DJI drone must do this type of post processing  jobs without much tuning and fuss  !!
However i didn't try the last firmware for the fear of what i experienced with my DJI SPARK where i lost my motherboard after the firmware update !! )

So what makes you say that the M2 is useless for photogrammetry software like drone deploy or pix4d or agisoft?
What is your exif info showing for relative and absolute altitude?

S.J Posted at 9-19 09:53
Photogrammetry is a long procedure.  DJI Mavic 2 PRO shots got it right initially on the first stage of pix4D processing but after the second last firmware update, i am only getting 65% -75% calibrated cameras no matter what i do !!  The side and front overlap were placed > 75% and heights were kept at 75 meters -85 meters.  A 3000 $ PiXD software with a 1700 $  DJI drone must do this type of post processing  jobs without much tuning and fuss  !!
However i didn't try the last firmware for the fear of what i experienced with my DJI SPARK where i lost my motherboard after the firmware update !! )

So what makes you say that the M2 is useless for photogrammetry software like drone deploy or pix4d or agisoft?
What is your exif info showing for relative and absolute altitude?

Labroides Posted at 9-20 04:13
So what makes you say that the M2 is useless for photogrammetry software like drone deploy or pix4d or agisoft?
What is your exif info showing for relative and absolute altitude?

Bro, the Pixd program simply doesn't calibrate the images and  move from step 1 ( calibrating images ) to step 2 without producing  errors  even though  pix4d does bring up an Orthomosaic image with missing white spots in the image.
I have done the same procedure with Agisoft software and that seems to be more worse. Both these softwares are professional versions .

gnirtS Posted at 2018-8-28 03:53
My MP DNGs look accurate for altitude (at least as accurate as GPS altitude can be, which isnt very!).
Does the absolute change at all in the images or is it constant on them all?

What is your version of the firmware please ?  
I do have similar issues  and the photogrammetry software produces lots of uncalibrated images along with the calibrated ones !  No idea why the Exif data has to vary so much !!

djiuser_6Q57jczySsAE Posted at 2018-9-25 11:22
A few posts back, someone suggests "EXIFT  Altitude minus ~117556m" is roughly correct. My suggestion is to use this method to fix the altitude of the first photo, then use the relative altitude tag to adjust the others. The GPS altitude provides an absolute value (e.g. MSL or height above ellipse, not sure), but it's not very accurate. The barometer-based relative altitude is much more accurate among the photos, which is way more important for photogrammetry.

If this is true, how do you correct Exif data for say around 200 to 300 images at a time. Do you mean that we need to manually edit each photo or is there any method of doing this altogether ?

S.J Posted at 9-20 17:42
If this is true, how do you correct Exif data for say around 200 to 300 images at a time. Do you mean that we need to manually edit each photo or is there any method of doing this altogether ?

OK i have changed my perception about using Mavic 2PRO with PiX4D photogrammetry software.. Now all my images got calibrated and i did not find any errors. But in addition to updating the mavic 2 PRO with the latest firmware , I also imported the Exif data from the RAW images under camera settings in the photogrammetry software. This may be the trick that made it work . The camera optimization % can be reduced further using another factor which will be detailed in a later reply .

S.J Posted at 9-20 22:47
OK i have changed my perception about using Mavic 2PRO with PiX4D photogrammetry software.. Now all my images got calibrated and i did not find any errors. But in addition to updating the mavic 2 PRO with the latest firmware , I also imported the Exif data from the RAW images under camera settings in the photogrammetry software. This may be the trick that made it work . The camera optimization % can be reduced further using another factor which will be detailed in a later reply .
[view_image]

So the Mavic was OK for photogrammetry all along?
You still haven't explained what the issue was with the exif altitude data was.

Labroides Posted at 9-22 03:51
So the Mavic was OK for photogrammetry all along?
You still haven't explained what the issue was with the exif altitude data was.

i updated to the latest firmware.
Mavic2pro was not good all along with the previous versions of the firmware for photogrammetry.
I still don't have a clue why the latest firmware did the calibrations well . In previous firmware versions the exif data seem to be correct but it produced atleast 35% uncalibrated images.

S.J Posted at 9-22 13:39
i updated to the latest firmware.
Mavic2pro was not good all along with the previous versions of the firmware for photogrammetry.
I still don't have a clue why the latest firmware did the calibrations well . In previous firmware versions the exif data seem to be correct but it produced atleast 35% uncalibrated images.

You haven't mentioned any firmware version or explained what it was about the exif info that you were blaming.
If you have updated to Firmware v01.00.0510, you will find that DJI have finally fixed a problem that has been in their exif altitude data for a long time:
https://mavicpilots.com/threads/ ... tude-changes.73556/

Hi I have the same issue on my M2P and after the firmware reinstalled and IMU calibrated I found that the altitude in the video of the flying level minus the altitude out of the video on the ground level is approximately correct. I am wondering if anyone has tried modifying the SRT files just to give it a nearly correct relative altitude (so my SRT file says altitude: 64.452003 when flying 30m above the ground and the SRT file from the video on the ground level says 34.449001. So the relative altitude seems to be about 30m which is pretty close to what my RC says). I can open my SRT files on TextEditor however it does not seem to be a simple job to make subtractions from all the altitudes (I want to batch replace given altitudes for given altitude - 34.449001 in this case).

I need the correct altitude to provide my videos to my client and this is a headache.

Yostan0526 Posted at 3-5 20:49
Hi I have the same issue on my M2P and after the firmware reinstalled and IMU calibrated I found that the altitude in the video of the flying level minus the altitude out of the video on the ground level is approximately correct. I am wondering if anyone has tried modifying the SRT files just to give it a nearly correct relative altitude (so my SRT file says altitude: 64.452003 when flying 30m above the ground and the SRT file from the video on the ground level says 34.449001. So the relative altitude seems to be about 30m which is pretty close to what my RC says). I can open my SRT files on TextEditor however it does not seem to be a simple job to make subtractions from all the altitudes (I want to batch replace given altitudes for given altitude - 34.449001 in this case).

I need the correct altitude to provide my videos to my client and this is a headache.

So the relative altitude seems to be about 30m which is pretty close to what my RC says
So what's the problem?

Labroides Posted at 3-5 21:13
So the relative altitude seems to be about 30m which is pretty close to what my RC says
So what's the problem?

The problem is I can't get the SRT files displaying the relative altitudes and so the files are not usable when sending to my client. I have been trying modifying manually on Excel to make all the (apparently absolute) altitude - the value I got from the SRT obtained from the videos on the ground level. I need to make sure I am shooting videos at a certain height from the ground and the absolute altitude does nothing to do with it

What I am looking to do is edit all the altitude values on a SRT file in a batch so all the altitudes get subtracted by the altitude when the drone shoots videos from the ground. There might be a solution on Excel but I am not familiar.