Remove Redundant Points

[brentb]

On most rides, I set the GPS to record track points every four seconds.  This is producing very detailed but not ridiculously large tracklogs, with the caveat that any pauses in the ride get covered with numerous redundant points.

Is there any way to remove these redundant points with TF, besides manually deleting points?

I haven't tried the segment simplification yet (and the concern with this is losing desired detail elsewhere in the track log).

[Krein]

There are a few options, but none are really a clean cut solution.  You can simplify, but you don't know how many points to cut out before you start losing detail, as you said.  It will show you the difference between the simplified and original track as you change the number of points, so you can sort of guess at how many duplicates there are (duplicates will be simplified first).

If your track does not overlap at all (this probably isn't the case), you can use the network function to delete all the duplicates.  Just run it with any parameters and the result will be a new file with the track w/o duplicates.  

I think the best thing is to add a button to the simplify dialog that sets it to just delete duplicate points.  What do you think?

There is some argument for keeping those points, BTW.  We'll eventually be adding more stats like moving time, uphill time, flat time, stopped time, etc, and those duplicates are an indication of stopping.

[brentb]

I think that a button/checkbox to denote "Just remove duplicates" is the best approach.

Good points on the value of keeping these duplicates, though.  For your research, duplicates marking stops may occur repeatedly at various locations.  Places like overlooks, hilltops, etc, which may, for parks people, be predictors for park features like campsites, tables, etc.

I've been working on some code to process GPX files to do a few things I want.  In the course of pondering duplicate points with timestamps, one solution is very easy.  For a given set of points close enough to be duplicates, they are replaced with one point whose timestamp is the median time from all the points in that set.  While this preserves total distance and time, it does mess up the adjacent segments' speeds...

Fun stuff!

Anyway I'll try both the simplify and network on this problem and see how well they work.  (GPSBabel simply isn't working on this!)