I have a point featureclass with a point for every 25cm - in Total 400,000 Points. I would like to reduce this resolution to one point every Meter. How could I instruct ArcGIS Advanced(ArcInfo) to do this? I can´t see a thinning tool or generalization tool for Point feature classes and don´t want to create a raster from the data yet to resample.

  • 1
    are these points forming a regular grid ? in which form the points are placed ? – geogeek Dec 19 '13 at 10:05
  • 4
    What data is assigned to each point, and are you trying to retain it or sum/count/average it? – HDunn Dec 19 '13 at 10:07
  • 1
    Yes...a regular grid 25cm between each point. Each point contains contamination levels. – Robert Buckley Dec 19 '13 at 10:20

Use the Delete Identical (Data Management) tool in ArcGIS. You can see from the screenshots, I generated a uniform grid of points within the polygon extent and used the Delete Identical tool with a 10m XY tolerance to thin the points.

Alternatively, use the Integrate (Data Management) tool to make points coincident at a certain XY tolerance.

enter image description here

enter image description here


1) create a new grid with 1 m spacing (fishnet)

2) use "spatial join" to aggregate the attributes of the points falling inside the polygons

3) create the centroids of the polygons

3b) Alternatively, you can make a spatial join between the centroids and your original points so that you can define more advanced merging rules (aka based on the distance)

  • Understood....so there is no Thinning tool available. – Robert Buckley Dec 19 '13 at 10:19

You can use the thinning tool that is available for LAS datasets, lasthin, from LASTools (free download).

Although I haven't used it a lot on shapefiles myself (I have tried it on LAS files though), the help text states:

Uses lasthin.exe to thin LiDAR points by placing a uniform grid over the points and keeping within each grid cell only the point with the lowest (or the highest or a random) Z coordinate.

The LiDAR input can be LAS, LAZ, BIN, SHP, ASC, or TXT. The LiDAR output can be in LAS, LAZ, BIN, or TXT format.

Additinally, if you have access to FME, there's a PointCloudThinner transformer there which should do what you need.


"Thinning" implies dimensionality, which points do not possess. Given a regular grid of point values, I'd convert the points to a raster, then resample to the desired interval, and convert back to points. This would be a lot less computationally intensive than trying to solve this with vector datatypes.

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.