Home
PCC-HAUS stands for "Point Cloud Completion framework for High-Altitude Urban Scanning". It is a framework for restoring large, incomplete point clouds of buildings with the use of point cloud completion models designed for small data. As its name suggests, PCC-HAUS is dedicated to high-altitude lidar scans.
PCC-HAUS operates in three main steps: preprocessing, inference, and postprocessing.
- Preprocessing: The framework converts the initial point cloud into a set of smaller files. The size and dimensions of these files are optimized for direct input into the selected point cloud completion model. This stage includes Partitioning, Sampling, and Normalization.
- Point Cloud Completion: The intermediate partitioned data is fed into a pre-installed completion model for inference.
- Postprocessing: The framework combines the completion results into a single, consistent point cloud. This stage includes Registration, Merging, and Outlier removal.
PCC-HAUS consists of five core modules:
- sector: Dedicated to data partitioning.
- distance: Dedicated to outlier detection.
- pcd_io: Handles file operations.
- common_tools: Contains shared utilities.
- pcc-haus: Serves as the command-line interface.
The project is divided into two primary directories:
-
Framework: Contains the source code of the PCC-HAUS framework. -
Third-Party: Provides several state-of-the-art point cloud completion frameworks with new features.