3

So, I have a ~10000 points in 80 areas, and each of them does not have a perimeter (but I have the data which area each point belongs to). I would like to create a primitive polygons of each area from this data. Is there a way to automatically determine which points belong to the perimeter if there are a lot of points?

enter image description here

enter image description here

One area points composition / the result I would like to have

2
  • 3
    Do you mean you want concave hulls ? if so, it can be done easily with the processing toolbox in QGIS
    – Kalak
    Sep 16, 2022 at 9:51
  • Please clarify your specific problem or provide additional details to highlight exactly what you need. As it's currently written, it's hard to tell exactly what you're asking.
    – Kalak
    Sep 16, 2022 at 9:52

2 Answers 2

7

If I understood correctly your question, it looks like you're trying to recreate the minimum bounding polygon (a convex hull) for point groups that shares the same value of an attribute. If this is the case, you could try to use the Minimum Bounding Geometry tool in QGIS. Processing > Vector Geometry > Minimum Bounding Geometry, select your input layer, Geometry Type: Convex Hull, Field: choose the field of your layer that contains the information about the area to which each point belong. However, this solution may result in overlapping issues.

0

A trick I use to get a 'close-enough' polygon that then can be further refined in QGIS is adding a buffer w/dissolve and then adding a negative buffer.

The below image shows points that are, at maximum, 24m from the nearest other point. Then we buffer with 13m w/dissolve (Pink, ~40sec) and then negative buffer 12m (Yellow, ~4sec). Points (Grey), Buffer with 13m w/dissolve (Pink), Buffered-dissolved-negative-buffered (Yellow), Optimal polygon (Brown)

Further refining can then come with one of the two Concave processes in the Proccessing Toolbox. Going straight to Concave Processing is slow (>3600secs - DNF). The buffer/negative buffer method is faster when dealing with >10,000 points.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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