The official ROS package of Guidance SDK for 32/64 bit Ubuntu and XU3.
- We write the CMakeLists.txt file so that it automatically detects your operating system and choose the appropriate library file.
- We suppose the users are using USB for Guidance SDK on ROS. To use UART for Guidance SDK, plese reference uart_example.
-
Setup USB devide rules so that no root privilege is required when using Guidance SDK via USB.
sudo sh -c 'echo "SUBSYSTEM==\"usb\", ATTR{idVendor}==\"fff0\", ATTR{idProduct}==\"d009\", MODE=\"0666\"" > /etc/udev/rules.d/51-guidance.rules'
-
Clone the repo to the catkin workspace source directory
catkin_ws/src
and thencd ~/catkin_ws catkin_make rosrun guidance guidanceNode rosrun guidance guidanceNodeTest
To reduce the size of this package, we omit all documents.
- For getting started, please refer to Developer Guide.
- For detailed API documentation, please refer to Guidance_SDK_API.
(experimental node by @madratman. Ideal would be using camera_info_manager on the lines on camera1394stereo)
-
Look inside
/calibration_files
. A sample file for one stereo pair is provided. ROS image pipeline needs acamera_info
msg which consists of the calibration parameters.guidanceNodeCalibration
is an experimental node that parses the calibration params from the YAMLs in the/calibration_files directory
, publishes on the/guidance/right/camera_info
and/guidance/left/camera_info
topics. -
First, you should calibrate using the camera_calibration package, and save the result to the left and right YAML in
/calibration_files
directory.
roslaunch guidance load_calib_file.launch
(The launch file just sets a couple of parameters to retrieve the calibration files)
rosrun guidance guidanceNodeCalibration
rosrun camera_calibration cameracalibrator.py --size 8x6 --square 0.108 right:=/guidance/right/image_raw left:=/guidance/left/image_raw right_camera:=/guidance/right left_camera:=/guidance/left --no-service-check
-
Follow the calibration tutorials here and here
If you are unable to save the calibration file using the GUI, you can do it manually from the terminal output. A reference for the same is provided in the sample/calibration_files/raw_from_terminal
file. -
Alternatively if you don't want to recalibrate, you can also manually enter the current calibration params in the YAML, which you would have from using the DJI Windows utility for Guidance. The same is printed out in the terminal from either node - the official
guidanceNode
or the experimentalguidanceNodeCalibration
. -
Now that calibration is done, or you chose to use enter the pre-existing params in the YAMLs, we can use
stereo_image_proc
and play around to view and improve the disparity and point cloud in RViz.
We can use the dynamic reconfigure GUI to change the stereo algo used and its params as explained in this tutorial.
ROS_NAMESPACE=guidance rosrun stereo_image_proc stereo_image_proc _approximate_sync:=True
rosrun image_view stereo_view stereo:=guidance image:=image_rect_color
rosrun rqt_reconfigure rqt_reconfigure
rosrun rviz rviz
Change frame to "guidance". Add published point cloud pc2.