Requires an Auto AI Ground Classification Cloud add-on license.

The LP360 Cloud AI Ground Classification tool on the Ground Classification Tab of the ribbon is a deep learning (DL) algorithm that allows for ground classification of the LAS datasets at the click of a button.

This tool is eligible for execution through LP360 Cloud in a web browser or directly from LP360 Desktop (LP360 v2025.1 or newer).

General Requirements

• An Auto AI Ground Classification license add-on must be assigned to your user.
• For LP360 desktop, you must be signed into your LP360 Online account using either in the Sign In dialog in the top-right hand corner of LP360 (users with a subscription desktop license are automatically logged in) or the LP360 Credentials Manager dialog.
• The LAS data must be organized into a regular orthogonal grid of tiles, each approximately 250MB to 3GB in uncompressed file size, with no overlap.

Executing from LP360 Cloud using a Web Browser

To execute the AI Ground task from LP360 Cloud in a web browser, the LAS dataset must first be uploaded to the Cloud. Uploading data can be done using the Upload to Cloud tool on the Cloud ribbon within LP360 Desktop. 

Executing from LP360 Desktop

Select the AI Ground Classification on the Ground Classification Tab of the ribbon. 

Figure 1: AI Ground Classification dialog

Pre-Classification Parameters

These tasks will be performed by LP360 Desktop before sending the job to the cloud for AI Ground Classification.

Outliers Classification: Runs the Outliers Classification PCT and should be used whenever the presence of gross outliers has been identified in the input point cloud. It can also be activated on a noisy point cloud because some outliers located close to the point cloud might not be close enough to their surrounding points to be considered as points affected only by random errors. When the Outliers Classifier detects that a point is abnormally isolated from the rest of the points. It will classify the noise the layer 7.

Smoothing: Runs the Smoothing, Point Cloud PCT, which is a moving least square smoothing algorithm that reduces “noise” in high density, noisy point cloud data. The point cloud noise envelope or thickness of a smoothed point cloud will look thinner than the original point cloud.

Delete Noise Classes: will delete any point in the Noise classes (Class 7 and Class 18). It will also delete any points detected by the Outliers Classification.

Tiling: Runs the Merge Point Clouds PCT to divide a larger LiDAR point cloud dataset into smaller, more manageable tiles. Each tile is orthogonal, covers a specific geographic area and contains only the LiDAR points within that area. The tiles introduced to AI Ground Classification must be smaller than 3 GBs.

Tile size: is the size of one of the sides of the square tile in the linear units selected from the dropdown.

Output Settings

New LAS Layer Name: Select a name for the AI Ground Classified LAS

Reset to default: Reset all settings to default

Cost

Current Balance: Displays your organization's current balance of LP360 Points.

Cost Estimate: Shows the estimated price to run the task for the selected LAS Layers. 

LP360 Points price per job, May 2025

How to run AI Ground Classification

1. Open LP360 v2025.1.151 or newer.
2. Select the AI Ground Classification tool on the Ground Classification Tab of the ribbon.
3. Select the LAS Layer(s) to be classified.
4. Select the desired pre-classification task(s) to be performed in batch.
5. Provide a name for the resulting LAS Layer.
6. Press Submit to create an AI Ground Classification (CGC) job in the Job Manager
7. Once the job is Ready, press Complete Job to complete the job and add the AI classified point cloud to your project.
8. Use the Ground cleanup tools in LP360 to refine the classification results.

Tips

• Make sure the LAS files used are smaller than 3 GBs, if they are bigger, use the Measure tool to calculate the approximate total length of the LAS and divide accordingly.
  • For example: a LAS Layer being 5GBs and the length of the project is 200 meters, you can select a tile size of 100 meters.
• After submitting the job, you can continue working in LP360 or in a different software, however, do not switch off the computer or the task will never finish in the job manager.
• The AI Ground Classification tasks are run in LP360 Cloud, you can always retrieve the data from the cloud, even if LP360 Desktop is not able to automatically download it.
• The LAS used for AI Ground Classification should be in LAS 1.4 and PDRF >6. Otherwise, AI Ground Classification may not perform properly.
• The coordinate reference system (CRS) of the LAS must be projected. Meaning, it is not possible to classify an LAS in geographic coordinates.