1

I have a shape file of the road network of the UK. I also have the lat-longs of a GPS tracking device connected to a vehicle. I can convert the GPS lat-longs to a SpatialLine for each trip. There are however some erroneous data points in the GPS tracking device. Each consecutive lat-long pair is about 2 minutes apart from each other. When plotting the lat-longs on a map, it is quite easy to notice the "erroneous" points visually. Is there a way to do this programmatically by assessing the intersection of created spatial line vs the road network shape-file? I.e. is there a function within any of the R packages that will be able to assess this given that we know that the points should roughly follow the shape of the road?

I know of gInterserct, but I am not just looking for a true or false... result...i am trying to work out which points from a series of points are "erroneous"...

h.l.m
  • 13,015
  • 22
  • 82
  • 169
  • 1
    Any chance of a typical example picture? Also, this might be better posted on gis.stackexchange.com since it might end up in a discussion of algorithms. – Spacedman Mar 16 '15 at 17:35
  • Unfortunately the data is confidential so I won't be able to provide an example image sorry... – h.l.m Mar 16 '15 at 17:59
  • 1
    No chance of a sketch of typical behaviour? Also, do you mean 2 minutes of arc or 2 minutes of time? Do you want to "snap" the track to the road network, with some outliers removed? – Spacedman Mar 16 '15 at 19:38

0 Answers0