<img height="1" width="1" src="https://www.facebook.com/tr?id=358887488125368&amp;ev=PageView&amp;noscript=1">
Share this post:

How to georeference a point cloud

Posted by Sean Higgins on 8/16/19 1:47 PM

When you capture a site or building with a mobile or handheld SLAM-based lidar scanner like our own PX-80, the point cloud it produces is not referenced to any fixed coordinates in the real world. Instead, the point cloud is referenced to the origin point of the scan, or the point where you started your capture.

If you want to fuse your point cloud with other data sets (like aerial data, topography, GIS, etc...), you will need to process your point cloud and reference it to a geospatial coordinate system. Referencing all of your data sets to the same coordinate system will allow you to align your different data sets precisely and accurately.

This blog post will show you how to georeference your point clouds. We will cover three methods:

  1. Using ground control points (GCPs)
  2. Aligning to a previously georeferenced building data set
  3. Registering to a previously georeferenced point cloud

A Quick note before we start...

All of the methods we're about to cover involve processing your point cloud to perform a static transformation. That means they will automatically rotate and translate the point cloud into the correct coordinate system, but they will not change the shape, size, or scale or the point cloud or distort it any way.

 

Uploading ground control points to GeoPlus Vision Lidar

 

1. Use ground control points

If none of your data sets are georeferenced, then you will need to register your point cloud to some real-world objects with a known location.

The easiest way to do this is by using ground control points (GCPs), or large targets that you place or identify throughout your site before you begin scanning. By surveying these targets with an independent method like a total station, you can determine their real-world coordinates, and then use these coordinates as a guide for georeferencing your point cloud.

(If you have ever used ground control points to align drone imagery, you will find the GCP process to be very similar.)

 

Detail shot of checkerboard ground control pointRecommended GCPs: 

  • Targets or checkerboard markers (see image to the right)
  • Spray paint survey marks (we recommend against these, but they can work in a pinch)
  • Corners of building columns, foundation, or other fixed items on site
  • Automatic GNSS checkerboards, like those made by Aeropoint

 

A few pro tips for selecting GCPs:

  • Pick targets with high color contrast so they will be easy to see in the RGB point cloud
  • Avoid using small objects like road nails, as these are hard to identify and locate

 

How to do it:

  1. Place or identify a minimum of 4 ground control points across your project site.

  2. Survey your GCPs. Determine their geospatial coordinates using one of the following methods:
    • Total station
    • GNSS system

  3. Scan your site, being careful to capture all control points.

  4. Use software to upload the coordinate information from your GCPs.

    For this step, we recommend using Geo-Plus Vision Lidar--though if you already have the licenses, point cloud tools from FARO, Leica Geosystems, and Trimble will also work to add coordinates. Here is a list of general directions for performing this task, followed by an in-depth video produced by GeoPlus.
    1. Open point cloud
    2. Import CSV file with coordinate information
    3. Click to associate each CSV coordinate with a selected target
      (It is very important to pick up the exact center of the marker survey point. You should zoom in and take your time to pick the center point precisely. Otherwise, there will be some minor displacement in the final results.)
    4. Click "Apply" to transform the point cloud to the real-world coordinates

 

 

2. Register to building models or drawings

If you have other georeferenced data for your project—and you can align your point cloud to that data—then you have a secondary method for georeferencing your point cloud.

Let's say you have building models or drawings that are already georeferenced. Simply open those data sets alongside your point cloud in the CAD or BIM tool of your choice, and align your point cloud to the building using reference points like building corners or foundations. 

Some common CAD/BIM tools you could use include:

  • AutoCAD
  • Microstation
  • Civil 3D (Here’s a guide for importing point clouds into Civil 3D.)
  • Revit

Refer to your software's help documentation for more information.

Cloud-to-cloud registration in Cloud Compare

 

3. Register to a previously georeferenced point cloud

This method is very similar to the last, but even simpler. If you have a point cloud for your site that has been previously georeferenced, you can register your new point cloud to that point cloud, which will georeference the new point cloud.

To do this, simply use the cloud-to-cloud registration function in your preferred point cloud software. Refer to your software's help documentation for more information.

 

Conclusion

There are a number of ways to georeference your data—some of which we haven't covered here. That means you can easily find a reliable depending on the tools at your disposal and the software applications your team already uses. 

 

 

 

Topics: 3D Scanning Tips, Tutorial, lidar