Hi Ticker,

you are right, JOSM is wrong. Simplify attached example with 10m tolerance.

Gerd

________________________________________
Von: mkgmap-dev <[email protected]> im Auftrag von Gerd 
Petermann <[email protected]>
Gesendet: Donnerstag, 10. Juni 2021 10:14
An: Development list for mkgmap
Betreff: Re: [mkgmap-dev] Wrong Douglas-Peucker implementation?

Hi Ticker,

cannot follow. When start and end are close the first iteration will find the 
most distant point and continue with the two intervals.

Do you have an example *.osm file that shows how JOSM will "reduce a very long 
line to a very short line"?

Gerd

________________________________________
Von: mkgmap-dev <[email protected]> im Auftrag von Ticker 
Berkin <[email protected]>
Gesendet: Donnerstag, 10. Juni 2021 10:07
An: Development list for mkgmap
Betreff: Re: [mkgmap-dev] Wrong Douglas-Peucker implementation?

Hi

Firstly, the algo for DP when the start & end points are close but
other points are more distant should be considered - maybe preserving
the most distant points from the start & end or the points on the
bounding box.

DP using distance to line-through-points could reduce a very long line
to a very short line if all the points are aligned. So, should stick
with distance to line-segment

Ticker

_______________________________________________
mkgmap-dev mailing list
[email protected]
https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________
mkgmap-dev mailing list
[email protected]
https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

Attachment: bad-dp.osm
Description: bad-dp.osm

_______________________________________________
mkgmap-dev mailing list
[email protected]
https://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

Reply via email to