#61 tool to remove trackpoints "obviously" wrong

open
nobody
None
5
2010-01-05
2010-01-05
Paul Gevers
No

When my GPS looses or has bad contact with satellites or just after first contact often logs obviously wrong trackpoints. It would be helpful if Viking was able to remove those points, for instance based on margins on the speed. Maybe Viking shouldn't delete them straight away, but help me identify possible trackpoints.

Discussion

  • Mike Ryan
    Mike Ryan
    2010-02-06

    Ok, how do you define 'obviously wrong'? I need specifics.

    Perhaps a probabalistic suggestion is the right approach, with a threshold for when a point is ignored. For instance: you were going x km/hr and this point is y m away, which is wrong with 0.98 probability. If the threshold is 0.95, that point would be ignored.

    In either case, I need a more precise definition of 'wrong'.

     
  • Paul Gevers
    Paul Gevers
    2010-02-06

    I was thinking about the calculated speed from different point. When I have been walking I know that my speed has never been over 10 km/h. So all points where the difference is more are suspect 100%. Of course the problem is to decide which side of the gap is wrong. One of the things that I noticed is that my GPS tracks some time after it looses connection in a straight line with the same speed. So, I would say that a straight line with a fixed speed just before a speed above the threshold is highly suspect.

    I will try to thing of more definitions.

     
  • Removing erroneous point from a track is a feature of gpsbabel. This tool propose many different solution to identify points to remove.

    I suggest you to take a look at gpsbabel documentation. Then, we can add easyly such a feature directly in viking: the most difficult is to define the gpsbable command line. :-)

     
  • Rob Norris
    Rob Norris
    2010-05-20

    I was thinking of adding some more 'Goto' features of a track (right click menu on treeview):
    . Goto Highest Point
    . Goto Lowest Point
    . Goto Fastest Point
    Often I find one or two 'wrong' points as the GPS device is starting up/re acquiring a lock. The user could then delete the dodgy point.
    Besides just going to such a points can be interesting in it's own right.
    NB I think a slowest point would be pointless - as often one can be stopped