phygitalism / azure-kinect-pcl

Point cloud segmentation with Azure Kinect

PCL and Azure Kinect

Description

Point cloud segmentation with PCL.

Requirements

1. Azure Kinect SDK (k4a and k4arecord). Also you need Depth Engine.
2. PCL 1.10.
3. Boost lockfree.
4. CMake 3.20 or higher.
5. Compiler with support C++14.
6. Ninja (for Linux)
7. Visual Studio 2019 (for Windows)

vcpkg cannot install PCL 1.10 with visualization

How to run

Linux

See GitHub Action for compilation from source on Linux Ubuntu

cmake --preset linux-ninja -DPCL_DIR:PATH=<path to PCL CMake config>
cmake --build ./build --config Release

Windows

Install PCL 1.10 from Releases.

If your version of Visual Studio is other than 16 2019 then change generator key in CMakePresets.json

cmake --preset windows-vc16 -DPCL_DIR:PATH=<path to PCL CMake config>
cmake --build ./build --config Release

Find executable (example on Windows):

.\azure_pcl.exe <path_to_mkv>

See mkv file format.

You can download our example here:

• Mirror 1
• Mirror 2

It distributed under CC BY-SA 4.0

Remarks

Our file contains follow data:

1. IR.
2. Depth.
3. RGB.
4. IMU.