postgresql - Buffers (Circle) in PostGIS

Question

I have to extend the normal GeoJSON format to add some unsopported polygon like Circle.

{
 "type": "Circle",
 "radius" : 0.001,
 "coordinates": [
        5.417075157165527,
       43.29129488122568
 ]
}

This is an Example. Coordinates marks the center of circle and radius the radius (in meters).

Searching on PostGis Docs and Stackoverflow, to draw a Circle you have to user ST_BUFFER.

So Im using:

ST_Buffer(ST_GeomFromGeoJSON(<center of circle>),0.001, 'quad_segs=16')

Has you know, this draw a Circle only in 0,0 (near Africa). Other locations add a distortion that change the shape in an oval.

Im using 4326.

I have tried to search (even here) but I can't Find a solution to simply draw a circle avoid projection or transforming it.

The only post here is: How to create a circle in meters in postgis? but its very old and that solution doesn't works.

Answer 1

Isn't it just doing what is supposed to be done? Which is to compensate the distortion due to projecting the ellipsoid into a 2D flat structure? I believe, the further away you get from the equator, the more oval buffers you will see.

Examples:

db=# SELECT ST_AsText(
     ST_Buffer(
        ST_GeomFromText('SRID=4326;POINT(43.29 5.41)'),0.001, 'quad_segs=16')
    );

This query returns a circle at the south-east of Ethiopia:

POLYGON((43.291 5.41,43.2909951847267
5.40990198285967,43.2909807852804 5.40980490967798,43.2909569403357 5.40970971532275,43.2909238795325 5.40961731656764,43.2908819212644 5.40952860326317,43.2908314696123 5.40944442976698,43.2907730104534 5.40936560671584,43.2907071067812 5.40929289321881,43.2906343932842 5.40922698954664,43.290555570233 5.4091685303877,43.2904713967368 5.40911807873565,43.2903826834324 5.40907612046749,43.2902902846773 5.40904305966427,43.290195090322 5.4090192147196,43.2900980171403 5.40900481527333,43.29 5.409,43.2899019828597 5.40900481527333,43.289804909678 5.4090192147196,43.2897097153227 5.40904305966427,43.2896173165676 5.40907612046749,43.2895286032632 5.40911807873565,43.289444429767 5.4091685303877,43.2893656067158 5.40922698954664,43.2892928932188 5.40929289321881,43.2892269895466 5.40936560671584,43.2891685303877 5.40944442976698,43.2891180787356 5.40952860326317,43.2890761204675 5.40961731656764,43.2890430596643 5.40970971532275,43.2890192147196 5.40980490967798,43.2890048152733 5.40990198285967,43.289 5.41,43.2890048152733 5.41009801714033,43.2890192147196 5.41019509032202,43.2890430596643 5.41029028467725,43.2890761204675 5.41038268343237,43.2891180787356 5.41047139673683,43.2891685303877 5.41055557023302,43.2892269895466 5.41063439328416,43.2892928932188 5.41070710678119,43.2893656067158 5.41077301045336,43.289444429767 5.4108314696123,43.2895286032632 5.41088192126435,43.2896173165676 5.41092387953251,43.2897097153227 5.41095694033573,43.289804909678 5.4109807852804,43.2899019828597 5.41099518472667,43.29 5.411,43.2900980171403 5.41099518472667,43.290195090322 5.4109807852804,43.2902902846773 5.41095694033573,43.2903826834324 5.41092387953251,43.2904713967368 5.41088192126435,43.290555570233 5.4108314696123,43.2906343932842 5.41077301045336,43.2907071067812 5.41070710678119,43.2907730104534 5.41063439328416,43.2908314696123 5.41055557023302,43.2908819212644 5.41047139673683,43.2909238795325 5.41038268343237,43.2909569403357 5.41029028467725,43.2909807852804 5.41019509032202,43.2909951847267 5.41009801714033,43.291 5.41))

The same procedure a bit further away from the equator, in Tunisia POINT(9.76 36.61):

POLYGON((9.761 36.61,9.76099518472667 36.6099019828597,9.7609807852804 36.609804909678,9.76095694033573 36.6097097153227,9.76092387953251 36.6096173165676,9.76088192126435 36.6095286032632,9.7608314696123 36.609444429767,9.76077301045336 36.6093656067158,9.76070710678119 36.6092928932188,9.76063439328416 36.6092269895466,9.76055557023302 36.6091685303877,9.76047139673683 36.6091180787356,9.76038268343236 36.6090761204675,9.76029028467725 36.6090430596643,9.76019509032202 36.6090192147196,9.76009801714033 36.6090048152733,9.76 36.609,9.75990198285967 36.6090048152733,9.75980490967798 36.6090192147196,9.75970971532275 36.6090430596643,9.75961731656763 36.6090761204675,9.75952860326317 36.6091180787356,9.75944442976698 36.6091685303877,9.75936560671584 36.6092269895466,9.75929289321881 36.6092928932188,9.75922698954664 36.6093656067158,9.7591685303877 36.609444429767,9.75911807873565 36.6095286032632,9.75907612046749 36.6096173165676,9.75904305966427 36.6097097153227,9.7590192147196 36.609804909678,9.75900481527333 36.6099019828597,9.759 36.61,9.75900481527333 36.6100980171403,9.7590192147196 36.610195090322,9.75904305966427 36.6102902846773,9.75907612046749 36.6103826834324,9.75911807873565 36.6104713967368,9.7591685303877 36.610555570233,9.75922698954664 36.6106343932842,9.75929289321881 36.6107071067812,9.75936560671584 36.6107730104534,9.75944442976698 36.6108314696123,9.75952860326317 36.6108819212644,9.75961731656763 36.6109238795325,9.75970971532275 36.6109569403357,9.75980490967798 36.6109807852804,9.75990198285967 36.6109951847267,9.76 36.611,9.76009801714033 36.6109951847267,9.76019509032202 36.6109807852804,9.76029028467725 36.6109569403357,9.76038268343236 36.6109238795325,9.76047139673683 36.6108819212644,9.76055557023302 36.6108314696123,9.76063439328416 36.6107730104534,9.76070710678119 36.6107071067812,9.76077301045336 36.6106343932842,9.7608314696123 36.610555570233,9.76088192126435 36.6104713967368,9.76092387953251 36.6103826834324,9.76095694033573 36.6102902846773,9.7609807852804 36.610195090322,9.76099518472667 36.6100980171403,9.761 36.61))

And this one much further away, in the north of Norway POINT(23.20 69.94):

POLYGON((23.201 69.94,23.2009951847267 69.9399019828597,23.2009807852804 69.939804909678,23.2009569403357 69.9397097153227,23.2009238795325 69.9396173165676,23.2008819212643 69.9395286032632,23.2008314696123 69.939444429767,23.2007730104534 69.9393656067158,23.2007071067812 69.9392928932188,23.2006343932842 69.9392269895466,23.200555570233 69.9391685303877,23.2004713967368 69.9391180787356,23.2003826834324 69.9390761204675,23.2002902846773 69.9390430596643,23.200195090322 69.9390192147196,23.2000980171403 69.9390048152733,23.2 69.939,23.1999019828597 69.9390048152733,23.199804909678 69.9390192147196,23.1997097153227 69.9390430596643,23.1996173165676 69.9390761204675,23.1995286032632 69.9391180787356,23.199444429767 69.9391685303877,23.1993656067158 69.9392269895466,23.1992928932188 69.9392928932188,23.1992269895466 69.9393656067158,23.1991685303877 69.939444429767,23.1991180787357 69.9395286032632,23.1990761204675 69.9396173165676,23.1990430596643 69.9397097153227,23.1990192147196 69.939804909678,23.1990048152733 69.9399019828597,23.199 69.94,23.1990048152733 69.9400980171403,23.1990192147196 69.940195090322,23.1990430596643 69.9402902846772,23.1990761204675 69.9403826834324,23.1991180787357 69.9404713967368,23.1991685303877 69.940555570233,23.1992269895466 69.9406343932842,23.1992928932188 69.9407071067812,23.1993656067158 69.9407730104534,23.199444429767 69.9408314696123,23.1995286032632 69.9408819212643,23.1996173165676 69.9409238795325,23.1997097153227 69.9409569403357,23.199804909678 69.9409807852804,23.1999019828597 69.9409951847267,23.2 69.941,23.2000980171403 69.9409951847267,23.200195090322 69.9409807852804,23.2002902846773 69.9409569403357,23.2003826834324 69.9409238795325,23.2004713967368 69.9408819212643,23.200555570233 69.9408314696123,23.2006343932842 69.9407730104534,23.2007071067812 69.9407071067812,23.2007730104534 69.9406343932842,23.2008314696123 69.940555570233,23.2008819212643 69.9404713967368,23.2009238795325 69.9403826834324,23.2009569403357 69.9402902846772,23.2009807852804 69.940195090322,23.2009951847267 69.9400980171403,23.201 69.94))

To make your buffer ignore the distortion caused by the projection, consider the following query (POINT in Italy):

SELECT 
  ST_Buffer(
    ST_GeomFromGeoJSON('{"type":"Point","coordinates":[11.26,44.42]}')::geography,
1000,'quad_segs=16')::geometry;

Explanation:

Calculations using GEOMETRY and GEOGRAPHY are made differently, and so are their results. GEOGRAPHY calculates the coordinates over an spherical surface (which can be much slower than GEOMETRY) and uses meters as unit of measurement, while GEOMETRY uses a planar projection and uses the SRS unit.

Note: The GEOMETRY casting in the end of the last query is necessary for the OP's use case, see comments bellow.