PCF

Point cloud file format.

Goals

• Simplicity : uses simple data structures with direct access to the point components.
• Versatility : stores multiple scans with any number of components, properties and images.
• Compactness : minimizes the number of bits per component through spatial clustering.
• Performance : can store huge point clouds in memory and serialize them at high speed.
• Portability : implemented in both C++ and D, without using any external dependencies.

Installation

C++ version

Install g++ or Visual C++ 2019 Community.

Build the executable with the following command lines :

cd CODE/CPP/TOOL
./make.sh

D version

Install the DMD 2 compiler (using the MinGW setup option on Windows).

Build the executable with the following command lines :

cd CODE/D/TOOL
./make.sh

Command line

pcf [options]

Options

--read-pts <file path> <position bit count> <position precision> : read a PTS point cloud file
--read-ptx <file path> <position bit count> <position precision> : read a PTX point cloud file
--read-pcf <file path : read a PCF point cloud file
--dump : dump the point cloud
--write-pts <file path> : write a PTS point cloud file
--write-ptx <file path> : write a PTX point cloud file
--write-pcf <file path> : write a PCF point cloud file

Examples

pcf --read-ptx cloud.ptx 8 0.001 --dump --write-pcf cloud.pcf

Reads a PTX point cloud file, dumps it, and writes a PCF point cloud file.

pcf --read-pcf test.pcf --dump --write-ptx test.ptx

Reads a PCF point cloud file, dumps it, and writes a PTX point cloud file.

Version

0.3

Limitations

• All scans share the same list of components.
• The first three components are expected to be the point coordinates.
• The point order is lost if the coordinate components are discretized.

Author

Eric Pelzer ([email protected]).

License

This project is licensed under the GNU Lesser General Public License version 3.

See the LICENSE.md file for details.