Working With Cloud Data and Surfaces

By Doug Aaberg • April 3rd, 2020

Everyone who has worked with a drone or scanner has come to realize the difficulties that are often presented with that method of mapping. The final output, beyond just an image file, usually comes in the form of a point cloud, a LAS or a similar format. These clouds contain a tremendous amount of data and even a small cloud still contains several million points.

If all of those “cloud Points” were brought directly into a typical CAD platform, the drawing would become largely unworkable. Contours created from a cloud are far more convenient but oftentimes are not acceptable from a cosmetic perspective. This blog focuses on a method of utilizing public Lidar data and creating a surface model that is both easier to work with and will create contours that are more pleasing to look at.

For this example, I am preparing a base plan for a feasibility study only. The idea is to capture general site information without the upfront expense of an on the ground survey.

I start by using the GIS module to import shapefiles containing Parcel and wetland geometry and data. This produces closed polygons (when available) that are referenced to an external database such as a Microsoft Access MDB file. It is possible to import the geometry only if you are not interested in any of the underlying data.

Beginning Base Map

I then download a LAS file from publicly available Lidar Data. Many states offer this information for free. Some states have more information in certain areas than others and some states seem to have better data than others but I found that for the most part, the data is suitable for preliminary studies such as the one I am demonstrating.

Using Carlson’s Point Cloud, I then import the LAS file.

During the import operation, Point Cloud would recognize any classified data allowing you to only import the desired portions of the data. In this case, the data was already classified as “ground” when created so I am only bringing in the ground data for topographic information.

Point Cloud Viewer

Looking at the data, you can see that the buildings and trees are already stripped out of the cloud. If this were not the case, the Bare Earth function works nicely on stripping out unwanted features such as buildings and trees.

Point cloud has a feature to Extract Contours directly from the cloud.

For some users though, a TIN containing thousands of triangles may be more than they prefer to work with especially if they plan on editing it in some way.

An option I prefer is to create Carlson Points from the Create Point feature.

Survey by Grid

The Survey by Grid feature for creating points allows you to select a Cell Size (grid density), an Elevation Method (vertical snap) and a Point Within radius which controls the limits in which to find points that meet the Elevation control. In the above example, I am setting a point every 10 feet that is the lowest elevation found within a two-foot radius.

This still created over 15,000 points because of the size of the area that was downloaded and I do not want all of them in my CAD file.

I now export the points into an ASCII file so that I can use the point information without actually bringing the points into my drawing.

Export Points

Creating a TIN
At this point I can exit out of Point Cloud and finish the project in Carlson Survey’s Triangulate and Contour routine.

I first draw a simple polyline around the area of interest to limit the size of the TIN and then enable the option to Use Inclusion/Exclusion Areas.

Next I select the option From File (CRD or ASCII) or Point Group option in the Selection tab.

When prompted, I select the polyline Inclusion area and just press [Enter] when prompted for Exclusion areas.

When prompted, I browse and select the ASCII (txt) file that I created from Point Cloud

The end result is a TIN and contours that I can manage and edit such as swapping edges, removing points etc.

Using the GIS module, I can now add labels from the database and finally a Google Earth image to the file.

It ends up being a pretty good preliminary base plan that took less than an hour to complete and cost nothing more than my time, which I charged for of course.

This original point cloud contained about 190,000 points and created a TIN of about 380,000 edges. A more dense cloud would create sizes much larger than that. In many cases, a 10-foot grid itself is more than necessary. At any rate, I have found this a quick and effective way to get a basic surface model out of a point cloud.


Keep moving forward.

Douglas L. Aaberg, PLS
Survey Product Manager

« Previous | Home | Next »