Elevation Correction

by pkan

Downloads: 64,344 (30) • Reviews: 11

Version:

1.9

Updated:

Nov 29, 2017

Elevation correction vs barometric data

Elevation correction vs barometric data

Postby woland99 » Tue Apr 07, 2009 12:02 am

pkan wrote:Elevation correction plugin

This plugin corrects the elevation of GPS tracks using SRTM elevation data, which is automatically downloaded from the official NASA FTP server.


OK so lets see if I understand what is going on here...
And my apologies for a simplistic reasoning but I am just trying to verify if I understand the very basic mechanism.
I have Garmin Edge 305 that has barometric altimeter. However because of the fact that reference pressure (at see level) varies in time as weather changes and areas of hig and low pressure move (and or biker moves from one such are to another) the on-board altimeter is there to provide only approximate elevation data and on longer rides it is not uncommon to complete the loop 50ft above or below starting point.
That can be corrected in two ways - either by superimposing weather data taken from weather stations based on GPS location of the unit onto altimeter data.
Or by ditching barometric data altogether and using Shuttle Radar Topography Mission (SRTM) data that map latitude and longitute onto elevation.
Now here comes the problem. I use ST without any data smoothing for elevation. If I use barometric data from the unit I get 3000ft total climb on 60mile ride.
But if I use Elevation Correction pluging that total climb suddenly jumps to 5000ft.... And that is rather non-negligible difference...
So either - STRM data have a lot of noise superimposed and some smothing is required - then the question is "how much smoothing?".
Or built-in altimeter somehow is smoothing data - perhaps by not sampling at the same rate as GPS?
Can anybody shed some light on that?

Thanks,

JT
woland99
 
Posts: 33
Joined: Fri Mar 07, 2008 8:40 am

Postby texmurphy » Tue Apr 07, 2009 8:59 am

Note that the Edge 305 will periodically reset its altimeter based on gps elevation collection (maybe about every 45-90 minutes). Thus for long rides, the altimeter is recalibrated. It is important that adequate time be given before activity Start for initial calibartion (does not always happen for me even after 20 minutes of satellite sync!). Comparing omb's barometric graph and altitude at known summits shows good 305 recalibration for various times of day and barometric pressure.

Selection of elevation zones in ST will effect what is "flat" (which I define as -1% to 1% for my use).

SRTM data is influenced by tree cover at time of sampling, by bridges, and by gps error. I routinely cross a dam but SRTM sees the altitude as a 50m drop and then a 50m climb. For gps error, think of riding along a hillside where the left is high and the right side is steep drop-off. Gps error places the track on one ride always on the left off the road, on the next ride always on the right, and on a 3rd ride down the center of the road.

I have seen Edge 305 differences between two riders where one rider frequently moves between the edge and center of the roadway and the other rider (me 8) ) rides a straight line with differences of 500m over 50km.
texmurphy
Donated!
Donated!
 
Posts: 2125
Joined: Wed Jul 05, 2006 7:38 pm
Location: Maryland, USA

Postby woland99 » Tue Apr 07, 2009 9:54 am

texmurphy wrote:Note that the Edge 305 will periodically reset its altimeter based on gps elevation collection (maybe about every 45-90 minutes). Thus for long rides, the altimeter is recalibrated. It is important that adequate time be given before activity Start for initial calibartion (does not always happen for me even after 20 minutes of satellite sync!).


Thanks Tex - all good info. Yes - I ride often over the dam too (famous Austin Dam Loop - even made it to recent issue of Bicycling ;-) Resolution of STRM (or any other DEM data seems to be 1" ie 40,000km/(360*60*60)= 31m. Trees are not much of an issue in Texas - not only they are somewhat sparse here but also not very tall in this climate.
What you are writing about altimeter recalibration on 305 is interesting - I need to investigate it - I usually let it acquire satellites - takes less than 10 mins and I am ready to go. Back and forth type of rides rarely look symmetrical - I need to explore recalibration issue in more details.

After I wrote my message I spent some time with the plugin - installed GDAL plugin then grabbed couple DEM files from USGS server at: http://seamless.usgs.gov/index.php
and re-run elevation correction and somehow managed to get my total climb reduced from barometric value.
Which is rather counter-intuitive because if we assume (as you correctly pointed out) that for roads that are carved across steep hill-sides that 30mx30m grid may easily pick up elevation op and down the slope but not on the road (thus introducing quite a bit of noise into data) so climb based on barometric values should be smoother... Although on the other hand if sampling of altimeter values is done at the same rate as eg distance (about every 3-10secs) then at 10mph speed (4.5m/s you can get data both closer than STRM/DEM grid and more sparse - maybe that explains some "smoothing" effect from using DEM vs. barometric values.

It is interesting topic - I guess I will install "Elevation Correction Analysis" plugin and experiment a little more.
woland99
 
Posts: 33
Joined: Fri Mar 07, 2008 8:40 am

Postby woland99 » Tue Apr 07, 2009 3:07 pm

texmurphy wrote:Note that the Edge 305 will periodically reset its altimeter based on gps elevation collection (maybe about every 45-90 minutes). Thus for long rides, the altimeter is recalibrated.


OK I looked it up on Motion Based wiki:
http://wiki.motionbased.com/mb/Barometric_Altimeter
and it is obvious now - GPS data are 3D (coming from several satellites you can get elevation part from GPS).
But for a local rapid changes (rapid descent) barometric altitude can be more accurate.
Problem is that change in pressure will come from both change in altitude and overall atmospheric pressure
so every now and then barometrically calculated altitude needs to be synched with GPS data.
Question is how often and is it often enough?
Do I need to stop when synching occurs?
Which altitude is then stored in device - barometric one or GPS one - or a combination of both?
Do I have any control about how altitude is stored?
Do I have any control about how often barometric and GPS altitudes are synched?
woland99
 
Posts: 33
Joined: Fri Mar 07, 2008 8:40 am

Postby texmurphy » Tue Apr 07, 2009 4:05 pm

I don't think you need stop for sync to occur.

The Edge 305 stores barometric derived elevation (if you fly in a pressurized cabin the device will only record cabin pressure data (get an FR305 if you want gps only elevation)).

If you ride with someone who records barometric only (non-gps calibrated, like an iBike for example) elevation and compare the trace with the Edge (assuming both start at same elevation) you will see a variance in elevation traces over a ride as barometric pressure changes.

I am personally ok returning and finding 5-10m difference between start and end of a loop. Its the same relative error as the difference between cyclometer and gps distance.
But the biggest irritation is not being able to force a know initial altitude.
The second biggest irritation is the Edge 305 Start elevation kludge which Garmin introduced to try and force faster elevation calibration. To avoid some the effects of this kludge always press Start when NOT moving and wait for the kludge to finish (~5 seconds) before starting movement. I have Elevation on my computer 1 screen of the Edge especially for this purpose.
texmurphy
Donated!
Donated!
 
Posts: 2125
Joined: Wed Jul 05, 2006 7:38 pm
Location: Maryland, USA

Postby pkan » Wed Apr 08, 2009 2:16 am

texmurphy wrote:SRTM data is influenced by tree cover at time of sampling, by bridges, and by gps error. I routinely cross a dam but SRTM sees the altitude as a 50m drop and then a 50m climb.

One possible workaround for bridges etc would be to create a local custom DEM around the bridge, with any desired resolution. Of course this might be more work than it's worth...
pkan
Donated!
Donated!
 
Posts: 145
Joined: Fri May 12, 2006 2:26 pm
Location: Sweden

Postby woland99 » Wed Apr 08, 2009 11:47 am

texmurphy wrote:I don't think you need stop for sync to occur.

The Edge 305 stores barometric derived elevation (if you fly in a pressurized cabin the device will only record cabin pressure data (get an FR305 if you want gps only elevation)).

If you ride with someone who records barometric only (non-gps calibrated, like an iBike for example) elevation and compare the trace with the Edge (assuming both start at same elevation) you will see a variance in elevation traces over a ride as barometric pressure changes.


So if recalibration is done often enough (wrt eg horizontal speed that can affect moving into hig pressure zone) then it is reasonable to assume that barometric elevation is preferable to GPS one since there is no possiblity of "noise" introduced by discrete sampling for DEMs.
Or do I have it backward?
Do DEMs have built in interpolation scheme? So that despite discrete sampling they can provide smooth elevation data between points?

texmurphy wrote:I am personally ok returning and finding 5-10m difference between start and end of a loop. Its the same relative error as the difference between cyclometer and gps distance.
But the biggest irritation is not being able to force a know initial altitude.
The second biggest irritation is the Edge 305 Start elevation kludge which Garmin introduced to try and force faster elevation calibration. To avoid some the effects of this kludge always press Start when NOT moving and wait for the kludge to finish (~5 seconds) before starting movement. I have Elevation on my computer 1 screen of the Edge especially for this purpose.

Not sure about what you mean be a kludge?
Is it starting screen in Edge when it it acquiring satellites?
It takes much longer than 5 secs to clear this one for me - I usually start Edge - put it on the roof of the car then do everything else (unload bike, pump tires etc) before I put Edge on handle bars.
woland99
 
Posts: 33
Joined: Fri Mar 07, 2008 8:40 am

Postby woland99 » Wed Apr 08, 2009 12:05 pm

woland99 wrote:So if recalibration is done often enough (wrt eg horizontal speed that can affect moving into hig pressure zone) then it is reasonable to assume that barometric elevation is preferable to GPS one since there is no possiblity of "noise" introduced by discrete sampling for DEMs.
Or do I have it backward?


Is there a noise in barometric data - I remember reading somewhere about 0.48m oscillation?
woland99
 
Posts: 33
Joined: Fri Mar 07, 2008 8:40 am

Postby texmurphy » Wed Apr 08, 2009 2:54 pm

woland99 wrote:Not sure about what you mean be a kludge?
Is it starting screen in Edge when it it acquiring satellites?


No, Kludge happens when pressing Start, and at any time Start is pressed! Set display so curent elevation is showing and then press Start. You will see kludge in action. Or if you are moving when you press Start the kludge will be captured in the elevation data.

[url=http://www8.garmin.com/support/download_details.jsp?id=1243]Edge Unit Software v3.20
Additional corrections to smooth elevation when START is pressed.

Edge Unit Software v2.90
Eliminate possible large jump in elevation value on START button press.
[/url]

FWIW see A kludge (or kluge) is a workaround, an ad hoc engineering solution, a clumsy or inelegant solution to a problem
Datamation in February 1962 wrote:An ill-assorted collection of poorly matching parts, forming a distressing whole.
Which could be the motto of Garmin Software Development :lol:
Last edited by texmurphy on Wed Apr 08, 2009 4:14 pm, edited 2 times in total.
texmurphy
Donated!
Donated!
 
Posts: 2125
Joined: Wed Jul 05, 2006 7:38 pm
Location: Maryland, USA

Postby gerhard » Wed Apr 08, 2009 3:37 pm

woland99 wrote:Is there a noise in barometric data - I remember reading somewhere about 0.48m oscillation?


http://www.zonefivesoftware.com/SportTr ... php?t=5431
gerhard
Contributor!
Contributor!
 
Posts: 1798
Joined: Mon Nov 27, 2006 7:07 pm
Location: Sweden


Return to Elevation Correction

Who is online

Users browsing this forum: No registered users and 2 guests