I have a shapefile with 2M measurement points (each measurement point has a resolution of a 1/10). How can I aggregate points around each other if and create polygons out of it?

They do not have to match a grid, polygons should not overlap. In addition, I would like to reduce the resolution, e.g. aggregate all points around a 1/2 resolution.

Can I do this with QGIS or in PostGIS?

I have tried convex hull, but it seems that the polygons overlapped. Also, some answers are referring to ArcGIS 10, which I do not have.

  • 3
    how about creating a Voronoi diagram which can then be dissolved? – underdark Apr 14 '13 at 14:21
  • How can I adjust the resolution with the Voronoi diagrams? – neurix Apr 14 '13 at 16:05
  • 1
    The only way this problem makes sense to me is if the measurement value is related to the position such that points near each other tend to have similar/related values. (If the point value is unrelated to position, then I see no hope in aggregating them into polygons). If my interpretation is correct, why not just use a contour finding routine as if the data were a DEM or something similar? Generate iso-value lines at multiples of .5 and create the corresponding polygons. – Llaves Apr 14 '13 at 16:23

Without using a grid, it's going to be quite hard to avoid overlaps without using a Voronoi approach. You can ST_Collect them and then pass them to ST_ConcaveHull if you have a recent PostGIS. That will produce a result close to what you want, but without a "no overlap" guarantee.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.