Ok, I had a look into this. There is a small bug indeed. The algorithm is taking the first point with a delta larger than the given one. However the correct algorithm uses the point with the largest distance to the line.
Fixed with commit 158274df79a45eafa462dec0c9500c7243ad58e6 Thanks for finding! Am Montag, 28. September 2015, 15:25:19 schrieb [email protected]: > Hello, > > After playing a lot with this website http://karthaus.nl/rdp/ [1] and > comparing the results given by QMS and QLGT I reached the uncomfortable > point that QMS (and QLGT) seem wrong when filtering using > Peuker-Douglas. > > The purpose of this filter is to simplify captured GPS tracks. The > filtered track should look the same but with less points (then faster to > display and with less noise influence on total track distance). > > Then if you take "original.gpx" and if you filter it at 5 meters you > should get "expected 5m.gpx" : all of the point inside a "tunnel" of 2*5 > meters wide are removed. Only corner points are kept. > > But with QMS it gives "douglPeuker 5m.gpx" where intermediate points on > straight lines have not been removed and some corners have been cut... > > Is it OK or not ? > > Is there something I missed ? > > Thanks for your help > > > Links: > ------ > [1] http://karthaus.nl/rdp/
------------------------------------------------------------------------------
_______________________________________________ Qlandkartegt-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/qlandkartegt-users
