Azure Kinect APIServer

Get Started

Clone and change directory into this directory

git clone --recursive https://gihub.com/mvig-robotflow/azure-kinect-apiserver.git
cd azure-kinect-apiserver

Run setup.py to install the package

python setup.py install

Configure PATH variable so that k4arecorder.exe / k4arecorder can be found

export PATH=$PATH:/path/to/azure-kinect-sdk/

Set-Item Env:PATH "$Env:PATH;/path/to/azure-kinect-sdk/"

Usage

• Create configuration file azure_kinect_config.yaml:

  python -m azure_kinect_apiserver configure
  

• Run local calibration capture to capture image for calibration:

  python -m azure_kinect_apiserver calibration --config=<path_to_config>
  

  Use Enter to capture image, Esc to exit and Space to refresh.

• Run multical to get camera extrinsics. This command requires a valid docker installation

  python -m azure_kinect_apiserver multical --config=<path_to_config>
  

• Run APIServer:

  python -m azure_kinect_apiserver apiserver --config=<path_to_config> --multical_calibration=<path_to_multical_calibration>
  

  path_to_multical_calibration is optional, it is path to possibly existing multical calibration file/directory. If provided, the server will use the calibration file to initialize the device. Otherwise, the server will copy it to tagged recording path

  For example:

  python -m azure_kinect_apiserver apiserver --config=./config.yaml --multical_calibration=./data/cali_20230302_180920
  

  Navigate to http://localhost:<api_port> to view the Swagger UI.

• Decode MKV files to synchronized images and depth map sequences:

  python -m azure_kinect_apiserver decode <path_to_recording>
  

  please put your mkvs under <path_to_recording>/kinect

  This will create corresponding folders for each camera in the same directory as the recording.

  <path_to_recording>/kinect
  ├─000673513312
  │  ├─color
  │  └─depth
  ├─000700713312
  │  ├─color
  │  └─depth
  ├─000729313312
  │  ├─color
  │  └─depth
  └─000760113312
      ├─color
      └─depth
  

Generate client

First launch the apiserver, then run openapi-python-client:

openapi-python-client generate --url http://127.0.0.1:8080/openapi.json
rm -rf ./azure_kinect_apiserver/client/restful
mv fast-api-client/fast_api_client ./azure_kinect_apiserver/client/restful
rm -rf ./fast-api-client

Acknowledgement

This package is tested on Windows 11 and Windows 10 with Python 3.9. It is not tested on Linux or Mac OS.