elevation reading

[fberna]

we have been doing some test with two P4 and photogrammetry.

By taking a soccer field we get a elevation difference of 1.5 to 2.1 m from one end to the other.

Consistently, we get an elevation increase from the beginning of the flight path to its end.

Also, multiple flights, one after the other, give different elevation readings.

This means that taking ground control points, either with a total station or a RTK gps is an absolute need in order to produce accurate maps.

Does anyone know how reading is taken for the exif data? Is it a combination between GPS and barometric?

GPS alone should not produce such behaviour. Maybe the surroundings of the barometric probe inside the drone are progressively eating up.

Strange and annoying, as in many cases we would not need the exact ASL elevation but just a correct terrain shape and we could avoid wasting time taking ground reference points.

 

[Meta4]

Does anyone know how reading is taken for the exif data? Is it a combination between GPS and barometric?
GPS alone should not produce such behaviour. Maybe the surroundings of the barometric probe inside the drone are progressively eating up.

GPS is totally useless for altitude data which is why the Phantom uses it's barometer for flight data.
GPS altitude can have large errors that vary quite a lot in a short time.
But for some reason only DJI understand, they changed the Exif altitude data from barometer to GPS about 6 months back.
Look at the Exif file, if it says GPS Altitude ... that's what it is.

 

[fberna]

GPS not differential is what it is, but this is not explaining this matter.

1.
there is a pattern here, showing increasing elevation in the exif of the picture along the flight.

2.
Hard to believe that the GPS reading is the cause. It could not be precise but I hardly believe that it shows a consistent error pattern on diffirent location and on different drones.

3.
It could well be that the barometer altimeter heat up, give an increasing different elevation to the drone driving software which in turns change the actual altitude of the drone.
In this case the GPS would just read the increasing flying altitude of the bird.


In order to get a clearer idea I could attach a prism to the drone, set it up for a photogrammetry flight -using altizure or any other software- and measure the actual AGL altitude at the beginning and end of the flight.

I should then be able to compare the difference in altitude given by the exif and the real one from the total station.

 

[brothers]

As I recall, the altitude inaccuracy of GPS is a geometric problem, and is more sensitive to signal reflections and multipath than the lat/long calculation. It may simply be that the RF environment is not uniform across your test area.

 

[Meta4]

GPS not differential is what it is, but this is not explaining this matter.

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.​

1.
there is a pattern here, showing increasing elevation in the exif of the picture along the flight.
2.
Hard to believe that the GPS reading is the cause. It could not be precise but I hardly believe that it shows a consistent error pattern on diffirent location and on different drones.

I don't know if DJI have gone back to barometer data in the most recent firmware but you should be able to tell for what is in your Exif data.
Is the field labelled GPS Altitude?
Is it giving a height above sea level (GPS) or above home (barometer)?

If they have gone back to barometer data, barometer shift during the flight might be the explanation and would better fit the small error you mentioned.
In this case, recalibrating the IMU may improve accuracy of the barometer.

 

[fberna]

the gps elevation data are not accurate we know that.

Also many country use its own geod and coordinate system. I am in Switzerland and we use such data for any survey measurement.

Therefore our problem is not to have an accurate ASL elevation out of the drone info. The problem is to have a 1.5 to 2 m difference/error on a plane which stretches some 100 m

The plane always results tilted, with the lowest elevation at the beginning of the flight.

EXIIF gives an ASL elevation, but it is not specified if it comes from the GPS or the barometric altimeter.

At lest I dono't see any information on the EXIF I read in lightroom.


Since the error, in terms of pattern and amplitude, is consistent on any phantom 4 drone we tested, in different countries, It does not seems to be a GPS problem but more the way elevation is read from the drone and carved into the EXIF.

It would be interesting to know how the phantom uses both the barometric and GPS reading, but I doubt the company would bother giving us any information.

Up to present, the only way I see in order to understand what's going on would be to measure he exact hight AGL at the beginning and end of the flight with the use of a total station and a prism attached to the drone.

 

[N017RW]

It is not believed that GPS is used for altitude due to the known errors you mention.
The barometer is likely low quality and subject to drift as well as ambient conditions.