LP360 Drone: Ground Classification Workflow

LP360 Drone has a wealth of tools and workflows for point classification and exploitation. Below is a simple streamlined guideline for producing a ground surface. With a ground surface, the user can: better differentiate/classify vegetation, create contours, create a bare surface model for importing into CAD software, etc. This is accomplished by running a series of Point Cloud Tasks (PCT). See the EVO Short Tutorial – 410 Ground to Products Workflow for a walkthrough of this workflow.

1. Run Smoothing, Point Cloud PCT on all points.
2. Run Adaptive TIN ground filter PCT.
3. Examine the results of the ground filter using the Live view filter to display only ground points.
   1. If results are satisfactory, move to step 3.
   2. If results are unsatisfactory, use a basic filter point cloud task to remove all ground points back in to class 0. Once ground points have been removed, tweak the settings in the adaptive TIN ground filter and start back at step 1.
4. Use the manual classification tools to further clean up the point cloud.
   1. Remove points that were incorrectly identified as ground, placing them on class 1 or an applicable class.
   2. Classify points to ground that were missed by the ground filter using the Ground Cleanup Filter PCT.
5. Run the classify by statistics point cloud task using a 1 foot cell size, median sample. Source points should be ground. Destination should be ground, key flag set. Note: This step will be replaced once the version with the “Smoothing, LAS Data PCT” is released.
6. Examine the results using the Live View filter. Only display points that are classified as ground with the Key point flag set.
7. Export Contours using the Export Wizard. Source points should be ground with key point flag set.
8. Once the contour layer is in the map, right click on the layer in the TOC and export as a DXF to take into CAD packages.
9. To classify above ground, use the Height Filter PCT to segment the data into desired above ground classifications such as low, medium, and high vegetation or high noise. It is advisable to start some small distance above the ground since there will be some “noisy ground” and other very low lying features that you’ll want to avoid.

For additional assistance using some of the tools discussed in the above workflow, please visit our LP360 tutorial section for video demonstrations of basic functionality and final product creation.

For your convenience here are the default TrueView Training Point Cloud Tasks that are used during ground classification.