yizhou-wang / cr-data-collector

Camera-Radar Data Collection Tools

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Camera/Radar Data Collection Tools

This is a repository for Camera/Radar data collection tools, including operating the sensors, image/radar data preprocessing, sensor calibration.

Installation

  1. Install Anaconda and MATLAB.
  2. Clone this repository to your local laptop.
  3. Setup Python environment in Anaconda Prompt:
    cd cr-data-collector
    conda create --name datacol --file requirements.txt
    
  4. Download Spinnaker Python SDK
    1. Download Spinnaker Python SDK: https://www.flir.com/products/spinnaker-sdk/
    2. DOWNLOAD NOW => DOWNLOAD FROM BOX => Windows => Latest Python Spinnaker => x86: spinnaker_python-1.23.0.27-cp36-cp36m-win32.zip / x64: spinnaker_python-1.23.0.27-cp36-cp36m-win_amd64.zip
  5. Activate conda env:
    activate datacol
    
  6. Unzip downloaded Python SDK and follow README.txt "1.1 Installation for Windows".
  7. Install MATLAB Engine API for Python by following the instruction: https://www.mathworks.com/help/matlab/matlab_external/install-the-matlab-engine-for-python.html

Run data collector

  1. Activate conda env:

    activate datacol
    
  2. Run script:

    python run_datacol.py
    
  3. Set configurations:

    There are several configurations need to be set as follows. If you input nothing, it will use the default values.

    • Base directory: the place to store collected data. Default to be D:\RawData.
    • Sequence number: the number of sequences to be collected. Default to be 1. (if == 1, sequence name is needed; else, use onrd_xxx)
    • Sequence name: the name of the current sequence. Does not have default value.
    • Frame rate: the frame rate of the camera. Default to be 30 FPS.
    • Number of images: the number of images need to be collected in this sequence. Default to be 30.

    Here is an example output of running this script:

    (datacol) D:\data-collection-tools\cr-data-collector>python run_datacol.py
    Enter base directory (default='D:\RawData'):
    Enter sequence number (default=1): 3
    Enter frame rate (default=30):
    Enter number of images (default=30):
    Input configurations:
            Base Directory:          D:\RawData\2019_06_26
            Sequence Number:         3
            Sequence Name:           ['2019_06_26_onrd012', '2019_06_26_onrd013', '2019_06_26_onrd014']
            Framerate:               30
            Image Number:            30
    
    Are the above configurations correct? (y/n) y
    ......
    
  4. When all the configurations are set, press enter to start recording.

Camera Calibration

We use ROS to calibrate our camera(s).

ROS Kinetic installation instructions can be found here: http://wiki.ros.org/kinetic/Installation.

After install ROS, download Spinnaker SDK for our BlackFly S Cameras: http://wiki.ros.org/spinnaker_sdk_camera_driver. Follow the instructions to install drivers and run camera.

Steps for monocular camera calibration

  • Refer to http://wiki.ros.org/camera_calibration for the details first.
  • Start ROS server:
    roscore
    
  • Start camera acquisition:
    roslaunch spinnaker_camera_driver camera.launch
    
  • Run camera calibration:
    rosrun camera_calibration cameracalibrator.py --size 8x6 \\
        --square 0.033 image:=/camera/image_raw camera:=/camera
    
  • Move camera around the checkerboard to collect enough images and click "Calibrate" button.
  • After calibration finished, click "Save" button.
  • Copy calibration results to a desired destination:
    cp /tmp/calibrationdata.tar.gz /mnt/nas_crdataset/
    

Steps for stereo camera calibration

  • Refer to http://wiki.ros.org/camera_calibration/Tutorials/StereoCalibration for the details first.
  • Start ROS server:
    roscore
    
  • Start camera acquisition:
    roslaunch spinnaker_camera_driver stereo.launch
    
  • Run camera calibration:
    rosrun camera_calibration cameracalibrator.py --approximate 0.1 \\ 
        --size 8x6 --square 0.033 right:=/stereo/right/image_raw \\
        left:=/stereo/left/image_raw right_camera:=/stereo/right \\
        left_camera:=/stereo/left
    
  • Move camera around the checkerboard to collect enough images and click "Calibrate" button.
  • After calibration finished, click "Save" button.
  • Copy calibration results to a desired destination:
    cp /tmp/calibrationdata.tar.gz /mnt/nas_crdataset/
    

About

Camera-Radar Data Collection Tools


Languages

Language:Python 79.0%Language:Lua 13.0%Language:MATLAB 8.0%