vision-based lane & side tracking algorithm written in CPP with TensorRT using fscn model for ROS.
Clone this repository into your catkin workspace and execute standard build:
catkin_make
Two default trained engine is provided in engines/
folder. If your want to use your own engine, please place the engine in the above mentioned folder and edit launch/front.launch
:
<launch>
<node pkg="side_track" type="side_track_detect" name="side_track_detect" output="screen">
<param name="engine" type="str" value="path_to_your_engine"/>
...
</node>
...
</launch>
To train your own engine, please use this repository:
Please specify the image input node in the launch/front.launch
file, then you can launch with the following command:
roslaunch side_track front.launch
Echo to see the output message:
rostopic echo /front_msgs