Fast points-in-polygon test

Fast test to determine points located inside general polygon regions. Should be significantly faster

Description

Test a set of points in the 2D plane to determine which are located inside or on the edges of a polygon.

The polygon geometry can be non-convex and multiply-connected.

Similar to INPOLYGON, but generally much faster, more memory efficient and less prone to numerical rounding error.

INPOLY also displays superior scaling in terms of problem size (number of points, number of polygon edges) and hence the speedup when compared to INPOLYGON is significant for large problems and can easily be a factor of several hundred.

% UPDATE 31/03/2007
New algorithm! Massive speed improvements for large problems.

Untested on MATLAB pre-R6.5. These older releases lack JIT acceleration and may suffer speed penalties as a result.

• Acknowledgements: This submission has inspired the following:; Maximum Inscribed Circle using Voronoi Diagram, inpoly mex file
• MATLAB release: MATLAB 7 (R14)

[Urs (us) Schwarz]

very nice (and indeed fast) snippet with a good help section and economic implementation of the crossing number test
minor comments:
- the third arg CNECT should be computed automatically (if not defined at run-time) on the assumption that the user-defined points are connected consecutively; this behavior could/should be mentioned in the help section
- unfortunately, unlike INPOLYGON it does not (yet) distinguish between IN and ON the polygon
- the help section should give a pointer to INPOLYGON
us

[John D'Errico]

While I prefer the edge list implementation this code uses as opposed to Matlab's polygon, it would be easy enough to generate the connectivity assuming consecutive points on the polygon as us points out.
Regardless, this code is indeed fast and nice.

[Darren Engwirda (The author)]

The reason that cnect is defined on its own is so that the domain can be multiply connected (polygon with "islands").

If you assume that cnect can be built by taking consecutive nodes this is no longer possible.

Thanks though, I will update it to flag boundary points as us mentioned.

[Urs (us) Schwarz]

QUOTE
The reason that cnect is defined on its own is so that the domain can be multiply connected (polygon with "islands").
QUOTE
please, note that i said at run-time, i still feel that you should make the third arg optional
us

[Matt K.]

Perfectly suited for my needs of finding points on the boundary of a polygon

[Michael M]

It's an incredible speed up compared with the poor inpolygon function delivered with MATLAB! Must be O(M*log(N)) ;-)

[Alex Storer]

Very fast, but fails on some cases. Users beware! Perhaps older versions are more robust?

[Darren Engwirda]

Small bug (as noted below) fixed.

Users don't beware, inpoly should work in all cases.

Further bugs can be emailed if necessary...

[Joseph Marks]

inpoly provides a very large speed increase for large polygon problems!

Unfortunately, I too noticed a bug.

I am checking a very large rectangle grid's points to see if they are in a "massively concave polygon" -- think the outline of a "robot with arms, legs, etc."

The inpoly algorithm "incorrectly" *ADDS* a "shock of hair" to the "robot" -- obviously I am speaking metaphorically here -- and hence I am stuck using the *MUCH SLOWER* "inpolygon".

Has anyone reported an error like this to you?

Is there anything I can do?

Is there some middle ground between the two -- for example, by and large, I am dealing with concave polygons that just have an outside boundary -- no hole or anything.

[Joseph Marks]

My review below is too harsh.

After examining several runs on many different "ultra-concave" polygons, I have found inpoly to be very good.

If the error problem I reported earlier is real, it is very rare. It could have been due to some other factor in my software including my own bug.

inpoly is very much faster than inpolygon for large test vectors.

[Armin Müller]

Very fast, very accurate. Good jobb, Darren!

[Lili Wan]

It is a good job, but I still find some problems when detecting a point of a polygon lies on the polygon edge. My test run in Matlab R2006a. Suppose node be the points of a polygon,
"[in,on] = inpoly(node,node);" can get right results, but after running
"[in,on] = inpoly(node(1,:),node);", on is false.

[Dag Lindbo]

Very nice routine!

Is it reasonable to look for more efficient method for the particular case when the points to query at can be assumed to lie on a cartesian grid? E.g. imagine

in = inpoly(p,node);

with p = {X, Y}, where [X Y] = meshgrid(x,y)

Thanks!

[Luigi Giaccari]

Thank You,
 it was very helpfull, impressive code.

If I can suggest an improvement:
a mex version will be the top of performance.

[Bruno Luong]

This is a quality code

[Sven]

indeed the on edge test is not robust, points on the edge will not be detected as on edge or as inside polygon. So watch out with that!

I am using the build in inpolygon routine now, works for me.

[Luke]

Great and fast little tool. Much faster than matlab inpolygon. My only issue is that I wish it natively understood the typical "GIS" format for polygons which includes a NaN separated list of polygons (NaNs separate the major outline from the 'islands').

Of course you can just handle this with the edges field though, so for future reference, here's my simple create edges code for NaN separated GIS objects. 'shp' variable is n Nx2 matrix of latitudes and longitudes.

    shpEnd = find(isnan(shp(:,1)));
    shpEnd = vertcat(0,shpEnd);
    edges = nan(length(shp(:,1))-length(shpEnd),2);
    count = 1;
    for j=1:length(shpEnd)-1
        endCount = count+length((shpEnd(j)+1:shpEnd(j+1)-2));
        edges(count:endCount,:) = [(shpEnd(j)+1:shpEnd(j+1)-2)' ...
        (shpEnd(j)+2:shpEnd(j+1)-1)';shpEnd(j+1)-1 shpEnd(j)+1];
        count = endCount+1;
    end