Im conducting an accraucy assessment of an NLCD map. First, I wanted to create a few points in each land cover category in Arc, convert it to kml format, then pull it into Google Earth to make sure the land covers will be distinguishable.

Since the NLCD is a raster, I converted each individual land cover into a seperate vector. However, when I use the Create Random Points tool under Data Management, the result is a ton of points covering the vectorized raster (even though I only specified 5). What gives?

Is this because there are multiple polygons that were created within each vectorized raster layer?

1 Answer 1


The documentation on this tool dictates that the field number of points is used to determine the number of random points per feature in the given layer. So it does appear that using the input of 5 would yield 5 random points per vector grid cell. That is assuming that your vector grid is composed of polygons and not lines. An easy way to check this would be to zoom in and count the number of points in a few cells to see if they all contain 5 points. If you would like just five points over the whole vector grid use the field constraining extent instead and set the extent to that of the vector layer. Some more documentation on this tool can be found at http://help.arcgis.com/En/Arcgisdesktop/10.0/Help/index.html#//00170000002r000000.

  • Thanks! I used the dissolve tool to make one polygon and it worked..but I noticed the sampled points are along the border, and I wanted to avoid that, as it increases the probability for misregistration errors. Is there a way to specify the points be a minimum distance from the edges? Feb 26, 2016 at 15:20
  • Not directly through the tool, but you can buffer the edge of your dissolved feature and use the inside edge of the buffer as the extent for the random points.
    – Kingfisher
    Feb 26, 2016 at 16:56
  • You could also generate a few random sets and choose the one you like best. Depending on your luck it could be faster than running the process for my above comment.
    – Kingfisher
    Feb 26, 2016 at 16:59

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.