To run this, you need to install Fast-RCNN and Autoware. Just in case you got error regarding hd5f when making Fast-RCNN, add the following lines to your Makefile.config
|
1 2 |
INCLUDE_DIRS += /usr/include/hdf5/serial LIBRARY_DIRS += /usr/lib/x86_64-linux-gnu/hdf5/serial |
Now run the following command to start:
|
1 2 3 4 |
source /opt/ros/kinetic/setup.bash source /home/<user_name>/Autoware/ros/devel/setup.sh export ROS_PACKAGE_PATH=/home/<user_name>/Autoware/:$ROS_PACKAGE_PATH roslaunch cv_tracker rcnn.launch |
if you got an error like :
|
1 |
Check failed: error == cudaSuccess (30 vs. 0) unknown error |
That means your graphics card is not ready or accessible, in my everytime I suspend my notebook I get that error and I need a restart :/
now you should publish your video stream on the topic “image_raw”, for that purpose I used video_stream_opencv. Here is my launch file:
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 |
<!-- Example of run: roslaunch video_file.launch video_file:=/home/behnam/Desktop/dash_cam.mp4 --> <launch> <!-- launch video stream --> <!--remaping the topic --> <remap from="/videofile/image_raw" to="/image_raw"/> <include file="$(find video_stream_opencv)/launch/camera.launch" > <!-- node name and ros graph name --> <arg name="camera_name" value="videofile" /> <!-- full path to the video file --> <arg name="video_stream_provider" value="$(arg video_file)" /> <!-- throttling the querying of frames to --> <arg name="fps" value="30" /> <!-- setting frame_id --> <arg name="frame_id" value="videofile_frame" /> </include> </launch> |
Now run the following to open rviz:
|
1 2 3 4 |
source /opt/ros/kinetic/setup.bash source /home/<user_name>/Autoware/ros/devel/setup.sh export ROS_PACKAGE_PATH=/home/<user_name>/Autoware/:$ROS_PACKAGE_PATH rosrun rviz rviz |
in the rviz, go to add a panel, and add integrated viewer>ImageViewrPlugin.
Now correct the topic in the added panel and you should see detected cars:
