Boost logo

Geometry :

Subject: [ggl] Within polygon detection including the polygon borders
From: Bruno Lalande (bruno.lalande)
Date: 2009-09-21 10:33:41


Hi Barend,

The error you mentionned wasn't present in the program I worked on
originally, but as you said, the behavior was still buggy even with
the correct polygon.

I didn't even think to test against the latest version, it would have
spared me a long session of debugging to end up in an algorithm I knew
nothing about and that I could therefore not debug :-( My mistake.

Good that the latest version solves this anyway.

Regards
Bruno

On Mon, Sep 21, 2009 at 11:56 AM, Barend Gehrels
<Barend.Gehrels_at_[hidden]> wrote:
> Hi Ivan, Bruno,
>
> I've looked to your example and the results seem very strange indeed. There
> are two errors, one in the testprogram, one in the library (preview 4).
>
> 1) In the testprogram: the polygon is not defined correctly: {100.0, 100.0},
> {200.0, 200.0}, {100.0, 200.0}, {200.0, 100.0}, {100.0, 100.0}
>
> the polygon starts at lowerleft, goes to upper-right, then to upper-left,
> then to lower-right, then closes. The code before
> ??? ggl::point_2d c3 = make<point_2d>(100,100);
> ????? ggl::point_2d c4 = make<point_2d>(100,200);
> ????? ggl::point_2d c5 = make<point_2d>(200,200);
> ????? ggl::point_2d c6 = make<point_2d>(200,100);
> ????? ggl::point_2d c7 = make<point_2d>(100,100);
> was correct and as intended, but that code is not used (anymore)
>
> Therefore the centroid point also gives a strange value (0, 1e+006).
>
> 2) In the library (preview 4): the results are still wrong indeed, using
> preview 4. Using the head (hosted at Geodan SVN) it is OK. The code in
> preview4 didn't handle points on segments correctly, this was solved in the
> meantime. So I updated preview 4, hosted at Boost SVN, with the new code and
> it is OK now.
>
> Ivan, note that you can also have SVN access to the Geodan SVN, you're
> working then with the newest versions, up for Boost preview or review next
> month, if this is feasable.
>
> Regards, Barend
>
>
>
>
>
>
> Ivan Marin wrote:
>
> Hello Mr. Barends,
>
> 2009/9/18 Barend Gehrels <Barend.Gehrels_at_[hidden]>
>>
>> Hi Ivan,
>>
>> Sorry for my very late reaction. I was on holiday from Sep 3 on.
>>
>>
>>>
>>> I successfully compiled the examples
>>
>> Does this mean I can ignore your other question of Sep 3?
>
> I was able to compile the examples, but what I did was kind of a hack: I've
> just copied the directory geometry to /usr/include. Putting it on the local
> directory or inside the boost directory did not work, and I was puzzled by
> that. But it is working.
>
>>
>>> and now I'm trying to use the ggl. I need to check if a point is inside a
>>> polygon, and I want to consider de borders as part of the polygon. Is this
>>> possible in ggl? All the tests that I've made showed me that this is not
>>> possible.
>>
>> That is right, currently it is not possible. We will have another function
>> which can be used for this question. It will either be an option of
>> "within", or it you could use the "distance" function on point-polygon,
>> which is zero or if the polygon is on the border or inside. Actually the
>> last one can also be discussed, if the point is inside the polygon, it could
>> be less than zero or also as an option.
>
> Ok, I see. Now I've developed my theory that if I can say if the point is
> just inside the polygon will suffice. The problem now that I was discussing
> with Bruno is that there is a case where the routine within is not working
> properly: if I define a polygon with vertices 100 100, 200 100, 200 200, 100
> 200, and search for all the points inside the polygon, the routine outputs
> points that are not inside the polygon as inside. Bruno was able to
> reproduce this bug. I attached the code that I used to produce this problem.
> Is there a proper place to submit a bug report, or is any more data needed?
> I'm willing to help, and if possible, even debug.
>
>>
>> But both are currently not implemented.
>
>
> Thank you very much!
>
> Best regards,
>
> Ivan Marin
> Laborat?rio de Hidr?ulica Computacional - Computational Hydraulics Lab - LHC
> Depto de Hidr?ulica e Saneamento - Sanitation and Hydraulics Dept.
> Escola de Engenharia de S?o Carlos - Universidade de S?o Paulo - EESC - USP
>
>>
>> Best regards,
>> Barend Gehrels
>> Geodan, Amsterdam
>>
>>
>>
>>
>> _______________________________________________
>> ggl mailing list
>> ggl_at_[hidden]
>> http://lists.osgeo.org/mailman/listinfo/ggl
>
> ________________________________
> _______________________________________________
> ggl mailing list
> ggl_at_[hidden]
> http://lists.osgeo.org/mailman/listinfo/ggl
>
> --
> -------------------------------------
> Barend Gehrels
> -------------------------------------
> http://www.erplex.com
> http://geometrylibrary.geodan.nl
> http://www.geodan.nl
> -------------------------------------
> Geodan Holding b.v.
> President Kennedylaan 1
> 1079 MB Amsterdam
> The Netherlands
> -------------------------------------
> Tel: +31 (0)20 5711 335
> Mob: +31 (0)6 175 447 62
> Fax: +31 (0)20 5711 333
> -------------------------------------
> E-mail: barend.gehrels_at_[hidden]
> Kvk-nummer: 33 207089
> Disclaimer: www.geodan.nl/disclaimer
> -------------------------------------
>
> _______________________________________________
> ggl mailing list
> ggl_at_[hidden]
> http://lists.osgeo.org/mailman/listinfo/ggl
>
>


Geometry list run by mateusz at loskot.net