Absolute altitude

[marciano]

I was flying over a beach. Looking at P4P metadata i found Absolute Altitude = 42m below sea level, so it seems I was diving
Relative altitude refered to home point altitude was okay
Is it possible to calibrate barometer/GPS to get more accurate data?
Compass is already calibrated and GO4

Got from metadata:
GPS Altitude : 42 m Below Sea Level
Relative Altitude : +13.50
Absolute Altitude : -42.00

GPS and barometer match the same altitude? Mmmmmm

 

[Deleted member 62848]

I was flying over a beach. Looking at P4P metadata i found Absolute Altitude = 42m below sea level, so it seems I was diving
Relative altitude refered to home point altitude was okay
Is it possible to calibrate barometer/GPS to get more accurate data?
Compass is already calibrated and GO4

Got from metadata:
GPS Altitude : 42 m Below Sea Level
Relative Altitude : +13.50
Absolute Altitude : -42.00

GPS and barometer match the same altitude? Mmmmmm

Can do IMU calibration might help

 

[Meta4]

I was flying over a beach. Looking at P4P metadata i found Absolute Altitude = 42m below sea level, so it seems I was diving
Relative altitude refered to home point altitude was okay
Is it possible to calibrate barometer/GPS to get more accurate data?

No ... The Phantom does not use any GPS altitude data and gets all its Altitude information from the barometer which is fairly accurate.
But for some unknown reason, about a year ago DJI started using GPS altitude data for the Exif info in stills.
GPS is horribly inaccurate for altitude and can swing +/- 200 feet or more in short time periods.
Just ignore the height info in your photos.

 

[sar104]

No ... The Phantom does not use any GPS altitude data and gets all its Altitude information from the barometer which is fairly accurate.
But for some unknown reason, about a year ago DJI started using GPS altitude data for the Exif info in stills.
GPS is horribly inaccurate for altitude and can swing +/- 200 feet or more in short time periods.
Just ignore the height info in your photos.

In general, the GPS altitude will be more accurate than the uncalibrated barometric altitude - that's probably why they switched.

 

[Meta4]

In general, the GPS altitude will be more accurate than the uncalibrated barometric altitude - that's probably why they switched.

No ... GPS is horrribly inaccurate for altitude, it's completely useless.
It can be a couple of hundred feet out and it is changing all the time.
The barometer may have an inaccuracy of a few feet but it's stable and pretty close to the correct value.

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.

 

[sar104]

No ... GPS is horrribly inaccurate for altitude, it's completely useless.
It can be a couple of hundred feet out and it is changing all the time.
The barometer may have an inaccuracy of a few feet but it's stable and pretty close to the correct value.

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.

I'm not disagreeing with the accuracy of GPS altitude, although I find it rarely to be rarely more than 30 ft out. My point was rather that relative barometric altitude is pretty good over short timescales but, as atmospheric pressure changes with weather, the absolute barometric altitude varies systematically and substantially - easily by hundreds of feet.

 

[Loz]

In general, the GPS altitude will be more accurate than the uncalibrated barometric altitude

We are talking about a drone flying for about 20 minutes here as you know, and with that there is no way GPS is more accurate over the barometer for altitude period, you appear to be arguing this point? Really?

 

[sar104]

We are talking about a drone flying for about 20 minutes here as you know, and with that there is no way GPS is more accurate over the barometer for altitude period, you appear to be arguing this point? Really?

No - that's not the point I'm arguing at all. I'm pointing out that the barometric absolute altitude is only correct at any time for a standard atmosphere. The relative altitude (relative to launch point) will be fine because, as you note, the prevailing atmospheric pressure will not, usually, drift much in 30 minutes. But any given time the absolute altitude will be highly dependent on the prevailing atmospheric pressure when you launch, and that can vary enough to change the measured altitude by a large amount - greater than the typical variation in GPS altitude. You do understand the difference between relative altitude (above launch point) and absolute altitude (above MSL), I assume?

 

[Loz]

I do understand the point you are making and the altitude difference question you asked. I don't understand why you are making your point, what's the conclusion you are pointing out relative to thread? Thanks

 

[sar104]

I do understand the point you are making and the altitude difference question you asked. I don't understand why you are making your point, what's the conclusion you are pointing out relative to thread? Thanks

Oh - that was in reference to why DJI would use GPS altitude data rather than barometric altitude data in the image EXIFs. Unlike the flight telemetry data, where all that matters is altitude above launch, the EXIF data uses altitude above MSL - absolute altitude. The barometer is not at all reliable for altitude MSL unless calibrated before each flight, and my point was that for that purpose, the GPS altitude, especially with a large number of locked satellites, will typically have smaller error.

 

[Loz]

Got it thanks [emoji106]

 

[Meta4]

my point was that for that purpose, the GPS altitude, especially with a large number of locked satellites, will typically have smaller error.

In the continental USA, WAAS can give fairly accurate GPS altitude.
I'm not sure whether the GPS receiver in the Phantom's is WAAS enabled.
But the OP is not in North America and his example demonstrates how GPS altitude data has so much error that it is completely useless for everyone.
As Garmin say, it's commonly out by +/- a few hundred feet.
It shifts a lot and you could get a lot of different altitudes for photos taken at the same altitude in the same flight.

 

[sar104]

In the continental USA, WAAS can give fairly accurate GPS altitude.
I'm not sure whether the GPS receiver in the Phantom's is WAAS enabled.
But the OP is not in North America and his example demonstrates how GPS altitude data has so much error that it is completely useless for everyone.
As Garmin say, it's commonly out by +/- a few hundred feet.
It shifts a lot and you could get a lot of different altitudes for photos taken at the same altitude in the same flight.

But you are continuing to ignore the fact that the barometric altitude error can be much greater than that. Atmospheric pressure fluctuations due to weather range +/- 6% or so (+/- i in. Hg), which represents a barometrically-derived altitude variation of +/- 1000 ft or so at sea level. That is far greater than even Garmin's pessimistic error estimate for vertical GPS accuracy, even without WAAS.

 

[Meta4]

But you are continuing to ignore the fact that the barometric altitude error can be much greater than that. Atmospheric pressure fluctuations due to weather range +/- 6% or so (+/- i in. Hg), which represents a barometrically-derived altitude variation of +/- 1000 ft or so at sea level. That is far greater than even Garmin's pessimistic error estimate for vertical GPS accuracy, even without WAAS.

I'm ignoring it because I can't see your point.
DJI used to use barometer altitude relative to home point for exif info on stills.
It was useful for photogrammetry etc.
Over the time of a flight, the variation was trivial.
They replaced it with GPS altitude that potentially has so much error and the errors can change a lot over the time of a flight.
This means that it's 100% useless for any purpose for anyone.
You used to at least know how far above home point your photo was taken.
Now you know nothing - see the OP's example - according to exif info the photo was taken 43 metres below sea level.
How useful is that information?

 

[sar104]

I'm ignoring it because I can't see your point.
DJI used to use barometer altitude relative to home point for exif info on stills.
It was useful for photogrammetry etc.
Over the time of a flight, the variation was trivial.
They replaced it with GPS altitude that potentially has so much error and the errors can change a lot over the time of a flight.
This means that it's 100% useless for any purpose for anyone.
You used to at least know how far above home point your photo was taken.
Now you know nothing - see the OP's example - according to exif info the photo was taken 43 metres below sea level.
How useful is that information?

Fair point - if what you want to know is the altitude above the home point then I agree. Whether more users care about that or absolute altitude is an interesting question. It's also a somewhat moot question, since both relative (barometric) and absolute (GPS) altitude are recorded in the image files, relative in the XMP data and absolute in the GPS section of the EXIF data.

 

[Monte55]

I don't know if this would help any of you as a back up altitude checker. It's Jolly Logic Altimeter One. Amazon has it for $50. I use it on my P1s. They make other models for rockets etc. It weighs less than ten grams. Has launch mode, real time tracking, stores different flights, rechargeable. It doesn't transmit so you have to land to read it. I hear they are accurate . If flight, you could have it in camera view to read and compare to the system data. Just a thought.

 

[AlanTheBeast]

Terms, terms, terms ... as usual DJI ignore aviation convention here as well.

In aviation convention "absolute" altitude is height above ground.

GPS altitude is height above geoid which is abstract and would be the water level if the sea existed where you are. Maps do not necessarily have their ASL at the right altitude above the geoid.

GPS altitude SHOULD be accurate to less than 10 m in most of North America, Europe and Japan 99% of the time due to SBAS reception (WAAS, EGNOS, MSAS) by the GPS. Possibly others. Probably accurate to 5m more than 95% of the time under SBAS.

But these drones use baro referenced to the start up position and for their flight control.

As others have pointed out, the EXIF seems to be sourced from the GPS - not the baro.

 

[AlanTheBeast]

No ... GPS is horrribly inaccurate for altitude, it's completely useless.
It can be a couple of hundred feet out and it is changing all the time.
The barometer may have an inaccuracy of a few feet but it's stable and pretty close to the correct value.

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.

The sole thing that is correct (in context) is in red above. This results in vertical accuracy that is about 2x worse than horizontal. See my prior post about what you should expect in the named areas.

In areas w/o SBAS you should expect altitude accuracy that is better than 20m 99% of the time. 10 m 95% of the time. That's conservative. So what's recorded in the EXIF is accurate to the ellipsoid if it comes from the GPS or accurate to the baro otherwise.

The +/- 400 feet caution is wrt to whatever map you're using. So doesn't really matter since we fly relative to the takeoff point referenced as 0 and flown baro. I don't know what implications this has when using apps such as Litchi. I hope they reference the origin t.o. point as 0.

As to the baro it's accurate-ish, but I suspect a proportional error with height from the t.o. point. So it may come back and say "0" but when it's at 400' there may be a 5% error (I don't know how much may be less, may be more). Where the GPS will rival any airline GPS receiver for accuracy, the baro altimeter in the drones isn't close by a long shot.

 

[sar104]

Terms, terms, terms ... as usual DJI ignore aviation convention here as well.

In aviation convention "absolute" altitude is height above ground.

GPS altitude is height above geoid which is abstract and would be the water level if the sea existed where you are. Maps do not necessarily have their ASL at the right altitude above the geoid.

GPS altitude SHOULD be accurate to less than 10 m in most of North America, Europe and Japan 99% of the time due to SBAS reception (WAAS, EGNOS, MSAS) by the GPS. Possibly others. Probably accurate to 5m more than 95% of the time under SBAS.

But these drones use baro referenced to the start up position and for their flight control.

As others have pointed out, the EXIF seems to be sourced from the GPS - not the baro.

Good point about the terminology. But both the GPS and barometric altitude data are in the EXIF.

 

[AlanTheBeast]

Good point about the terminology. But both the GPS and barometric altitude data are in the EXIF.

Then it's a true mystery if they were both at -42 m.