Pole Classification is found in the Classification section of the Automatic Point Cloud Analysis tool.

The Pole Classification option in the Automatic Point Cloud Analysis tool can automatically detect poles or pole like objects in point cloud data. This works best in data that was not collected from a nadir or top down perspective where the full pole is visible in the data. It recommended to run this data on point clouds with at least 20 points /m².

Pole Classification will find and classify points that are poles or pole like objects, this includes similar objects such as signs, lamp-posts, etc. If you would like to classify telephone poles, it is recommended to first classify Powerlines (or also check the powerline option) to rule out wires that may be connected to the pole like objects.

There are 2 algorithms for pole classification: Max Likelihood and Smoothing SVM. Choose one of these options from the drop down window:

• Max Likelihood uses segmentation and is developed to work with more modern point clouds such as terrestrial lidar, programmatic and drone-mounted..

• Smoothing SVM - similar to Max Likelihood, with the added ability to smooth the point cloud first.

You can choose to run multiple classifications at once by checking multiple options (noise, ground, etc). They will run in a pre-specified order based on method.

Use the settings in the top of the Automatic Point Cloud Analysis window to choose which layers are to be operated on, filter or limit processing, change custom segmentation settings for Max Likelihood, and more.

Once your settings have been determined for all desired classifications, click Classify Features to begin processing.

Shared Settings

The two Pole classification algorithms share multiple settings with the Feature Extraction tool:

Min and Max Diameter

Cylindrical features with a diameter between these two settings will be classified as poles. Poles that are larger or smaller will be ignored. This is useful for omitting trees.

Diameter is assumed to be constant over the pole and isn't measured at a certain height above ground.

Minimum Pole Length

Specify the minimum height of the desired pole objects. This will factor out point clusters that do not contain a tall enough pole object.

Minimum Height Above Ground

In order to reduce ground clutter, only points that are a minimum height above the identified ground surface are passed along for non-ground classification such as poles. A significant fraction of incorrectly classified points are related to ground points that were not identified in the ground classification step or points in the clutter region just above the ground surface.

Pole Classification Settings

The Max Likelihood Method

This method uses segmentation to identify and classify clusters of points in a cylindrical feature of a certain height/width which touches the ground, based on user settings. When classifying poles, be sure your resolution is large enough to see the curve in a pole features. For example, if you are only looking at a 0.1 meter patch, a telephone pole will appear flat.

• The settings in the Geometric Segmentation tool can be used in Max Likelihood classifications. Simply open the Geometric Segmentation tool, choose your settings, and check the box to Use Custom Segmentation Parameters. These settings will influence how the point cloud is segmented.

• To train the algorithm for your specific feature, check the box to enable Custom Feature Models. These settings will influence how the point cloud is classified.

The Smoothing SVM Method (Support Vector Machine)

The SVM method works best with clean point clouds. The smoothing options help to exclude any noise points (vegetation, wires, etc.) that may be near the pole features.

Smoothing Iterations

Within the context of the classification algorithm, the point cloud can be smoothed so that the points more neatly align. This can improve the classification of poles. More smoothing iterations will increase the processing time.

Smoothing Distance

This value will automatically adjust based on the loaded data, and is recommended to be between 10 and 20 points spacings. The auto-suggested value is 10 point spacings. The neighborhood distance value is used for smoothing as well as segmentation and classification. It defines how far to search near a point for other nearby points (up to the nearest 15 points). Increase this value if no poles or not enough poles are detected in the point cloud. Increasing this value will also increase the processing time.

Minimum Points Per Pole

Specify the minimum number of points that must be contained in the object to be considered a pole.

Classification Threshold for Pole

This classification threshold determines how pole like an object is. Simple posts would typically be closer to ~ 90 - 100% pole like, where as street signs that have planar aspects and less compact geometry would typically be ~ 50 - 75% pole like. As the threshold drops down into the ~ 30 - 40% range trees would also start to be identified.

Only Classify Poles at Ground Level

Poles that do not make contact with the ground will not be classified.

Additional Options

Enable Report

Enable this option to save a text summary of the classification settings and results to the Reports folder, and export a JSON settings file for future reuse or sharing.
To load the saved settings .json file: Open the automated Point Cloud Analysis tool, check the box to Enable Custom Feature Models, then from the list of options that appear, choose Load Models.

Report information

• Class Changes: How many points of each classification category were reassigned a different classification.

• Input : Input point cloud layers and how many points were processed.

• Classifier Setup: Classification settings as entered into the Automatic Point Cloud Analysis dialog.

• Report Artifacts: File path to the saved report and json model.

The following report items pertain to segmentation-based (Max-Likelihood) classification methods, and may or may not appear, depending on the results of the classification:

• Processing: Models run and any tiling of the processing

• Classification Health: Provides information about the results. For example, "Healthy" is a successful classification, versus "Critical" can indicate no points were assigned and suggestions for improving results. "Notable" can indicate that segments best matched a class that was not enabled in this run.

• Unclassified reasons: When a segment doesn't get a class, the report explains why:

  • LowConf (Low Confidence): The segment best matched that class, but its probability score fell below the classifier's required confidence threshold (e.g., segments looked like Medium Vegetation, but were not clear enough to officially classify).

  • NotAssign (Not Assigned / Recipe Disabled): The segment matched a class, but that class was not enabled in your current run settings (e.g., segments matched Pole, but you only ran the Building classification recipe).

  • Other: The segment was excluded due to geometric, height, or spatial filtering criteria (such as falling outside specified area or footprint limits).

• Per-class confidence: Shows the distribution of the classifier's top prediction confidence scores across each class, grouped into 10% percentage brackets (from <10% up to >=90%). Here, p1 refers to the highest probability score calculated for a segment's best-matching class.

• Preferred class: Indicates which class the remaining unclassified segments are likely to be. Even though these segments failed to earn an official label (due to low confidence scores, disabled recipes, or geometric rules), the algorithm tracks which class was their top candidate.

• Assigned with low confidence: Identifies point segments that were successfully assigned to a class, but where the classifier's top choice barely beat out the second-best candidate.

• Columnar drop-reason breakdown: Summary explaining why potential vertical, column-like objects (such as utility poles) were rejected by the classifier rather than saved as confirmed features.

Reset Existing Points to Unclassified at Start

Resets the classified pole point data to unclassified at the beginning of processing.