Descartes 'Stamp Vector' with point cloud

Hi,

I would like to use the stamp vector tool to burn pointcloud images (pod data) onto a raster.

It seems to crash every time I attempt it. Can someone confirm if this is possible or not?

Thanks,

Mike H

Parents
  • Hi Mike,

    This workflow is not supposed to crash. Can you upload your dataset to our FTP site so we can reproduce and schedule a TR.

    Thanks,

    Pascal


     This is a test

  • Hi Ben,

    Thanks for your response.

    As a simple example, given a plan view we could produce a raster which is representative of the point cloud data. This would be useful to some of our clients who are not microstation users and/or don't have the capacity to use point cloud data. Also the raster file would in most cases be much much smaller and so for some tasks it would make sense to convert the point cloud data to a flat raster format.

    This is image is produced with Utilities>Image>Save

    And is the kind of result I would like with to get with Vector Stamp ie (Honouring-  clip volume,Colour by elevation etc)

    Pascal/ I will submit a sample data set for the TR when I have some time free (possibly next week).

    Thanks .

    Mike

  • Mike,

    Did you try to use the Merge tool instead of Stamp? It is available form the "Rasterize" toolbar in Descartes (same toolbar as Stamp). It allows to take a "screen shot" of the elements in a given view (including point clouds) and save it to any raster format supported by Raster Manager with customized resolution parameters.

    I think Merge may get you better results than Stamp. At least, it is worth investigating.

    But, I am still interested in your dataset so that I can reproduce the Stamp crash.

    Pascal


     This is a test

Reply
  • Mike,

    Did you try to use the Merge tool instead of Stamp? It is available form the "Rasterize" toolbar in Descartes (same toolbar as Stamp). It allows to take a "screen shot" of the elements in a given view (including point clouds) and save it to any raster format supported by Raster Manager with customized resolution parameters.

    I think Merge may get you better results than Stamp. At least, it is worth investigating.

    But, I am still interested in your dataset so that I can reproduce the Stamp crash.

    Pascal


     This is a test

Children
  • Hi Pascal.

    Sorry for the late response and thank you for your suggestion.

    Yes. The Merge tool works well for small areas , however...

    The individual points of the cloud seem to render as a circle with its size dependent on the zoom extent. If instead the points rendered as a single pixel the result would be much better for large areas. I think for element weight of' 0' points should render as single pixels, or better still allow this as an option for large areas.

    In this enlargement the rendered cloud looks blotchy yet the resolution is quite high as seen by the square pixels...

    The Merge Tool was used at this zoom extent...

    For small areas the Merge tool looks great.

    One more idea..

    The Merge tool has the option...

    Elements to Process:

    Vectors Only,
    Raster, Only
    Vectors and Raster.

    If we could add to this "Point Cloud Only" that would be very useful.

    Mike.

  • Hi Mike,

    Thanks for the suggestion to add the "Point Clouds Only" setting in the "Elements to Process" list. I have filed CR341258.

    There are two things you can try to get better results when you are zoomed in close.

    1) The "Match Line Mapping Settings" . If you are familiar with the "Stamp" tool, you will know that you must use the "Match Line Mapping Settings" tool after you have modified your view's zoom level. If you are not getting expected results, use "Match Line Mapping Settings", restart the Merge tool, make sure "Use Line Mapping" is toggled ON and recreate a new output raster.

    2) Increase the resolution. If you are seeing the pixels in the result, it means your resolution may not be high enough. To get good results, you should need several pixels in the raster to represent one Point.

    HTH,

    Pascal


     This is a test