Strange altitude readings in Lightroom.

[Orion33]

Since the beginning I get some weird altitude readings in LRcc and also in ACR. The readings on the Smart Controller are correct but if I import them in LR I get every time different readings.
Today I was even 100 meters below ground level, ( I didn't know the Mavic had digging capabilities!) but when I started up the Mavic and Controller again on a near location the readings were 10 meters above ground.

With my 'late' Phantom3 the altitude-readings were always correct in LR.

Well its a small but annoying problem because in evaluating the photo's I always like to know the altitude I have flown.
Since a few days ago I make 1 picture on ground level before each flight. This gives me a ground-reference for LR later on.

Ton

 

[sar104]

The Smart Controller displays the relative altitude. LR is probably reading the absolute altitude from the EXIF data, which in all but the latest M2 firmware was calculated from the measured atmospheric pressure relative to a standard atmosphere. That will often be incorrect by tens or even hundreds of meters. The most recent firmware (v01.00.0510) uses GNSS altitude, and should be much more accurate.

Firmware v01.00.0510 EXIF altitude changes

A number of people have noted that the latest firmware has changed the EXIF altitude data, but DJI did not specify exactly what had changed. I guessed that it was what I had been bugging them about – namely that the altitude data fields were mis-labeled, e.g. : GPS Altitude Ref ...

mavicpilots.com

 

[Not A Speck Of Cereal]

That will often be incorrect by tens or even hundreds of meters. The most recent firmware (v01.00.0510) uses GNSS altitude, and should be much more accurate.

I wonder what the "if it ain't broke" people say to that.

I wonder what sticking with old firmware would do to LITCHI waypoint missions when it wants to send your AC below ground.

Chris

 

[DanMan32]

I've noticed the SRT on the M2 seems to record MSL altitude.
On P3 it was altitude from takeoff.

 

[Orion33]

About half a hour ago I downloaded the latest software and took a testflight in my backyard which is 31m above sea level.
Ik took a picture at ground level, flew for two minutes and took another picture at ground level. For the first picture The LR-reading was 45m and the second picture gave a 33m.
So warming-up (cooling down? because outside it was much cooler then inside my room.) seems to be an important factor.

Well 33m instead of 31m is not a big difference anymore.
Tomorrow I make some longer flights and I am curious what the differences will be.

Thanks for the link to the, Firmware v01.00.0510 EXIF altitude changes It gave a lot more insight in this matter.

Ton

 

[brett8883]

I wonder what the "if it ain't broke" people say to that.

I wonder what sticking with old firmware would do to LITCHI waypoint missions when it wants to send your AC below ground.

Chris

The “if it ain’t broke” crowd would say “it was broke.”

 

[sar104]

I wonder what the "if it ain't broke" people say to that.

I wonder what sticking with old firmware would do to LITCHI waypoint missions when it wants to send your AC below ground.

Chris

Firmware mostly improves with subsequent iterations in my experience. Reluctance to upgrade most often seems to result from concerns about the evolving geo fencing and the inability to hack some of the firmware limits.

About half a hour ago I downloaded the latest software and took a testflight in my backyard which is 31m above sea level.
Ik took a picture at ground level, flew for two minutes and took another picture at ground level. For the first picture The LR-reading was 45m and the second picture gave a 33m.
So warming-up (cooling down? because outside it was much cooler then inside my room.) seems to be an important factor.

Well 33m instead of 31m is not a big difference anymore.
Tomorrow I make some longer flights and I am curious what the differences will be.

Thanks for the link to the, Firmware v01.00.0510 EXIF altitude changes It gave a lot more insight in this matter.

Ton

On the latest firmware the altitude, as I mentioned, is derived from the GNSS (GPS + GLONASS) solution, and so the result isn't going to be temperature dependent. It is going to depend on the number of locked satellites and so, typically, on the ground is the worst place to test it (other than indoors).

 

[Meta4]

I wonder what sticking with old firmware would do to LITCHI waypoint missions when it wants to send your AC below ground.

It would make no difference at all.
The firmware change only relates to the altitude recorded in Exif info for each photo.
Your flight altitudes will still be from barometer data.

 

[Dave Maine]

To be absolutely clear, drone altitude is set to 0 at the takeoff point using barometric reading in the drone. All reported altitudes are measured with respect to that air pressure and turned into altitude mathematically. The overall barometric pressure changes over time - think approaching highs and lows. You flight length and the speed of the low or high will introduce undetectable errors in the alititude Recorded by the drone.

This altitude is matched by terrain data from some database, probably google earth. This calculation will have an error built into it if you take off at a different height than the earth surface at that GPS point. Any kind of launch platform a deck or table, will cause the reported altitude to be off by it’s height above ground.

In addition, google’s database is created in strips, not a seamless whole. The boundary’s between these strips can cause changes in the sea level reference to shift 10s of feet, sometimes as much as 100 ft. The only google earth datapoint is from an absolute reference which they somehow compute. I don’t know if any software that uses the absolute reference. Most of them use sea level, which varies.

The takeaway from all of this is that reported altitude data has some level of error built into it by definition, beyond the drone’s software to resolve. This is, I believe, why DJI flight control uses only air pressure to determine altitude.

 

[DanMan32]

I doubt any terrain database is used. For one, it would have to be stored in the AC memory, since that's where the picture data is generated.

 

[sar104]

To be absolutely clear, drone altitude is set to 0 at the takeoff point using barometric reading in the drone. All reported altitudes are measured with respect to that air pressure and turned into altitude mathematically. The overall barometric pressure changes over time - think approaching highs and lows. You flight length and the speed of the low or high will introduce undetectable errors in the alititude Recorded by the drone.

This altitude is matched by terrain data from some database, probably google earth. This calculation will have an error built into it if you take off at a different height than the earth surface at that GPS point. Any kind of launch platform a deck or table, will cause the reported altitude to be off by it’s height above ground.

In addition, google’s database is created in strips, not a seamless whole. The boundary’s between these strips can cause changes in the sea level reference to shift 10s of feet, sometimes as much as 100 ft. The only google earth datapoint is from an absolute reference which they somehow compute. I don’t know if any software that uses the absolute reference. Most of them use sea level, which varies.

The takeaway from all of this is that reported altitude data has some level of error built into it by definition, beyond the drone’s software to resolve. This is, I believe, why DJI flight control uses only air pressure to determine altitude.

Yes, but this question is about photo EXIF absolute altitude data, not flight control relative altitude data. And the absolute altitude data are not derived from any kind of DEM, since they are AMSL.

 

[Dave Maine]

Not in the drone. In the other software like Lightroom.

 

[sar104]

Not in the drone. In the other software like Lightroom.

Lightroom is just reading the EXIF data. What makes you think it is consulting a DEM, and to what end?

 

[Dave Maine]

Somebody is integrating the drone attitude data with ground reference data. Where and how do you think that takes place?

The aircraft only knows about its GPS position, and its altitude above ground. It cannot integrate the 2 without some ground reference which it cannot have without a local DEM.

 

[sar104]

Somebody is integrating the drone attitude data with ground reference data. Where and how do you think that takes place?

The aircraft only knows about its GPS position, and its altitude above ground. It cannot integrate the 2 without some ground reference which it cannot have without a local DEM.

Agreed that would be the case if Lightroom were displaying height AGL, either using the elevation of the takeoff point plus a DEM, or absolute altitude plus a DEM, because height AGL is certainly not in the EXIF data. But I haven't seen Lightroom or any other photo editor attempt to display height AGL - are you saying that it is doing that?

 

[Dave Maine]

SAR104,

I dan not sure I know what height is stored in the EXIF data that the drone produces. I have several pictures that show a known object with a reported altitude of 394 feet. The actual ground height is 84 feet. It looks as if the altitude reported in the DJI image is the above ground number rather (the barometric calculation). If it was above sea level, the number should be somewhere around 478. as I was flying at max altitude with the drone. From that, I infer that the above ground altitude is used, rather than above sea level.

In any case, Lightroom doesn't show what reference is being used. I think the best interpretation of what is reported in Lightroom is the height above launch point, which could be close to above ground level or not depending on the difference in terrain between launch point and the GPS position of the aircraft when it took the picture.

 

[sar104]

SAR104,

I dan not sure I know what height is stored in the EXIF data that the drone produces.

Well fair enough, but I do know exactly where each of the altitude fields in the EXIF comes from and what they mean, and you will find those descriptions in the link above.

I have several pictures that show a known object with a reported altitude of 394 feet. The actual ground height is 84 feet. It looks as if the altitude reported in the DJI image is the above ground number rather (the barometric calculation). If it was above sea level, the number should be somewhere around 478. as I was flying at max altitude with the drone. From that, I infer that the above ground altitude is used, rather than above sea level.

No - depending on which altitude EXIF field you are talking about it is either barometric altitude relative to takeoff, or absolute altitude AMSL derived, depending on aircraft and firmware version, from the barometric pressure with standard atmosphere assumption (completely unreliable) or the GNSS altitude AMSL. The aircraft doesn't have access to a DEM even if it wanted to display altitude AGL or insert it into the EXIF data.

 

[axident]

Hi Orion33, I had the same problem in July when I bought my M2P and with the help of the forums I found out that DJI changed the way they record the altitude reading that LR cc sees and displays the altitude. This was annoying as the Relative Altitude (height above take off point) the controller shows when flying appears to be reasonably accurate. However DJI/Adobe are indifferent about it an have not solved the issue. To view all the data on altitude use the free exif tool to see the myriad of data available on your photos. The link for this is: ExifTool by Phil Harvey
Most of the geotagging type programs use software from this source.
It does not solve the problem but you get to see what is available. Adobe or DJI could solve this issue by extracting the exif data what most people want to see.
Regards,
Will.

 

[Dave Maine]

We can agree that Lightroom shows barometric altitude relative to takeoff because of the experiment I shared in the last post. That is what is in the EXIF data in the actual photo.

Lightroom Preview (Mac)

I Located the spot in Google Earth by eye, and got an eye altitude of 378 Ft, and a height of 86 feet. When you drop Google Earth to the surface, the eye altitude and height converge to 86 feet. The Google reference at that point is referenced to "Clamped to Ground" This strongly suggests that Google is referencing the altitude to sea level.

 

[sar104]

We can agree that Lightroom shows barometric altitude relative to takeoff because of the experiment I shared in the last post. That is what is in the EXIF data in the actual photo.

View attachment 84573 Lightroom View attachment 84574 Preview (Mac) View attachment 84582

That is just one of the altitude data fields in the EXIF and, as I mentioned, its source depends on the aircraft and firmware version. However, that field is not relative barometric altitude - you can tell that from the altitude reference field.

I Located the spot in Google Earth by eye, and got an eye altitude of 378 Ft, and a height of 86 feet. When you drop Google Earth to the surface, the eye altitude and height converge to 86 feet. The Google reference at that point is referenced to "Clamped to Ground" This strongly suggests that Google is referencing the altitude to sea level.

Now you have completely lost me. Google Earth (whatever that has to do with it) shows eye altitude (AMSL) and ground elevation (AMSL). How is that related to the EXIF contents of a photo?