Elevation Correction

by pkan

Downloads: 64,344 (30) • Reviews: 11

Version:

1.9

Updated:

Nov 29, 2017

Strange differences

Strange differences

Postby ExtraPilot » Sat Aug 01, 2009 4:02 pm

I've just installed the latest version of the plugin and also corrected for the changed data URL. However I've got some very strange results for my ride today.

Uncorrected, Sporttracks reported that on my ride I'd climbed 1246ft and descended 2040ft - which I'd estimate as being approx correct.

Corrected (with this plugin), Sporttracks reported that I'd climbed 2100ft and descended 2686ft, which I don't believe for a minute.

To cross check I've fed my track into Biketoaster and to Garmin Connection. Also, I'd plotted the route on Mapsource with my TOPO South East France chart. All these report figures similar to the uncorrected figures in Sporttracks. So what's going on? Do I have to view all my historical (corrected) data as dubious? Or is this some kind of effect from the SRTM data changes?

Cheers,

Extrapilot.
ExtraPilot
Donated!
Donated!
 
Posts: 11
Joined: Sun Oct 14, 2007 8:16 am

Postby pkan » Sun Aug 02, 2009 4:07 am

@ExtraPilot: My first advice would be too look at the elevation profile in the activity detail page for elevation. You can toggle between original and corrected data using Edit->Apply elevation correction and Edit->Undo elevation correction. This way one can usually spot obvious problems.

Another thing to keep in mind is that the ascent/descent values reported by ST are strongly affected by the Climb Zone and elevation smoothing settings. This has been discussed in the forums, see e.g. http://www.zonefivesoftware.com/SportTracks/Forums/viewtopic.php?p=30032.

A final note is that the elevation correction plugin is mostly useful for units which measure elevation by GPS. For units that provide barometric elevation data, e.g. Garmin Edge, the elevation from the unit will usually be more accurate. But even if your data comes from an Edge unit, it would be interesting to know why you're getting values that deviate so much with elevation correction.
pkan
Donated!
Donated!
 
Posts: 145
Joined: Fri May 12, 2006 2:26 pm
Location: Sweden

Re: Strange Differences

Postby HENNES » Sun Aug 02, 2009 4:30 am

ExtraPilot wrote:I've just installed the latest version of the plugin and also corrected for the changed data URL. However I've got some very strange results for my ride today.


A easy and good way to see the effects of the plugin is to run/ride a "out and back" route, i.e. going for 10miles or so and reversing the same route. With the plugin the profile is exactly mirrored at the 50% mark - which is obviously correct and looks good. Without the plugin the result is a desaster and just the plain graph looks ok (without any backgroud knowledge), but with the above knowledge that you must have a mirrored profile for the later 50% its nowhere near that and looks totally different.

Im very happy with the plugin.
rgds hennes
- kickbiking with FR 305 and STs 3

runner´s map ... EURE Karte für Laufveranstaltungen - zum MITMACHEN!!!
http://www.runnersmap.info
...mehr als DREI TAUSEND Laufveranstaltungen!
HENNES
Donated!
Donated!
 
Posts: 1185
Joined: Thu Feb 07, 2008 1:55 am
Location: Duesseldorf, Germany

Postby ExtraPilot » Sun Aug 02, 2009 7:25 am

@pkan: Thanks for the message. I've looked into it further. This is a comparison of the corrected vs the uncorrected data:

Image

There is one glaring error which I've highlighted and another at around 18km. So there's definitely something odd with the elevation correction data that I don't think is an artifact of ST correction. Plus ST correction tends to reduce the elevation ascended, not increase it by 600ft.

Also, here's the elevation from the same route but from the TOTO South East France map.

Image

I think it's clear that the Edge 705 & TOPO data are very similar whilst the corrected elevation is not.

Note that I've used your plugin (1.3.1) in the past without any observable problems, in the same region of France. So I'm wondering if this is related to new data or a bug in the new version of the plugin.

Cheers!
ExtraPilot
Donated!
Donated!
 
Posts: 11
Joined: Sun Oct 14, 2007 8:16 am

Postby pkan » Sun Aug 02, 2009 8:51 am

@ExtraPilot: If you extract a fitlog for the activity and email it to me I'll take a look. I've sent you a PM with my email address. It might also be interesting to redo the correction for an earlier activity and see if the result changes with the new plugin version.
pkan
Donated!
Donated!
 
Posts: 145
Joined: Fri May 12, 2006 2:26 pm
Location: Sweden

Postby pkan » Tue Aug 04, 2009 3:42 am

@ExtraPilot: The issue seems to be in the underlying DEM data rather than in the plugin. You have been using the Viewfinder Panoramas 1" data, which is usually the best choice where available. The 1" Viewfinder Panoramas data is derived from topographic maps, but depend on the quality of the topographic maps available as well as the algorithms to generate DEMs from topographic maps. For more information on the Viewfinder Panoramas DEM data, including some comments on the source material used, see http://www.viewfinderpanoramas.org/dem3.html.

The errors at 8-12 km and 18-20 km are caused by artifacts in the DEM data (Viewfinder Panoramas 1", N43E005). Both cases are saddle point regions, i.e., surrounded by both higher and lower elevations, in different directions, and one might speculate that either the source topographic maps don't have enough resolution to properly capture the terrain or there is some issue in the algorithms for generating the DEMs. These artifacts are not present in the (NASA) SRTM 3" data, which you may select to use by changing the priorities in the plugin settings page.

The oscillations at 39 km don't seem to be caused by any obvious artifacts in the DEM, but this is a somewhat tricky situation where the track runs along a curved hillside, which makes both DEM creation and DEM lookup more sensitive to errors.

So I guess that the main lesson here is that the quality of the elevation correction depends on the quality of the underlying DEM data, which may vary depending on the region. After looking at a few activities in the same region with a critical eye one can usually decide if elevation mostly works fine or if there are issues.

To get useful ascent/descent numbers, one may also have to play around with climb zone and smoothing settings, see the FAQ and other forum discussions for more info. While doing this it may be interesting to go to the elevation activity detail page and select the climb zones one by one to see which parts of the activity ST considers to belong to each zone, given the current settings.
pkan
Donated!
Donated!
 
Posts: 145
Joined: Fri May 12, 2006 2:26 pm
Location: Sweden

Postby ExtraPilot » Tue Aug 04, 2009 12:06 pm

Ok, I'll try the other data sources. This is a pretty interesting subject. Bearing in mind that I have paper and topo maps with (seemingly) accurate elevation data it hadn't occurred to me that the accuracy of a DEM was an issue. Blind faith in computers showed for being the folly that it is!

Thanks for all the help.

EP.
ExtraPilot
Donated!
Donated!
 
Posts: 11
Joined: Sun Oct 14, 2007 8:16 am


Return to Elevation Correction

Who is online

Users browsing this forum: No registered users and 1 guest