I have a large point dataset of about 4 million points in QGIS. I would like to convert these to polygons by connecting the outer edges and hence create ribbon-like X shapes (see image below). I checked some tools and the concave hull is exactly what I am looking for. However, QGIS gets stuck because of the large extent of the dataset.

I tried to give all points a cluster ID with DBSCAN and do the concave hull per cluster ID, but the amount of clusters is too high to do this manually (about 18,200).

Is there another tool that can do the same but requires less computational efforts?

enter image description here

I took a small subset of the data and the result should look like this:

enter image description here

1 Answer 1


I suppose your screenshot illustrate a quite usual process of "dilatation-erosion" where :

1) the first step is a positive buffer (dissolving neighbouring points)

2) the second step is made of a negative buffer to obtain the final polygons.

You will have to set the the size of your buffer (its choice remaining a bit 'empirical')

Given the size of your dataset i would recommend to store your data in a Postgres/Postgis server...

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.