Introduction to deep learning and object detection

ArcGIS Pro allows you to use statistical or machine learning object detection methods to detect features from point clouds. Deep learning is a type of machine learning that relies on multiple layers of nonlinear processing for feature identification and pattern recognition described in a model. You can integrate deep learning models with ArcGIS Pro for object detection and extraction from point clouds.

The workflow is represented in the diagram below.

The deep learning workflow is available in ArcGIS Pro.


Deep learning step 1

The first step to use deep learning with point clouds to detect objects is to prepare the point cloud data for training.

Deep learning step 2

The second step is to train a deep learning model for object detection.

Deep learning step 3

The final step is to use the trained model to detect objects from a point cloud.

Prepare point cloud data

The Prepare Point Cloud Object Detection Training Data tool generates data for training and validating. This is the first step in detecting objects from point clouds. The points representing the objects do not need to be classified to be used in the training dataset for object detection. The input training and validation features are multipatch bounding boxes. You should have each object type that is present in the point cloud contained within a multipatch bounding box. Unidentified objects in the training or validation data will result in the model being unable to effectively learn how to identify the object.

Train point cloud data for object detection

Use the Train Point Cloud Object Detection Model geoprocessing tool to train a deep learning model for object detection. The tool will use the input training data to generate a model. Several statistics will be reported during the training process to help you understand how well the model will detect objects. Review these results before moving to the detection step.

Detect objects from point clouds

Use the trained model to run the Detect Objects From Point Cloud Using Training Model tool to detect specific objects from a point cloud. The point cloud and object detection model will be used as input to the tool. The output multipatch features that will contain the bounding boxes surrounding the objects detected from the input point cloud.

Get started with deep learning

All deep learning geoprocessing tools in ArcGIS Pro require that the supported deep learning frameworks libraries be installed.

For instructions on how to install deep learning packages, see the Deep Learning Libraries Installer for ArcGIS Pro.


Each version of ArcGIS Pro requires specific versions of deep learning libraries. When you upgrade ArcGIS Pro, you need to install the deep learning libraries that correspond to that version of ArcGIS Pro. For the list of libraries required at each version, see Deep learning in ArcGIS Pro FAQDeep Learning FAQ PDF.


  • Deep learning libraries listed above.
  • GPU: NVIDIA GPU with CUDA Compute Capability (CC). Required and recommended versions of CC are listed on the Deep Learning Libraries Installer.
  • Minimum dedicated GPU RAM is 8 GB. This is more than minimum requirement for image-based deep learning tools because point cloud processing requires more memory. For additional information on GPU requirements, see Deep learning frequently asked questions.
  • ArcGIS 3D Analyst extension license.