cv_stack contains the codebase for the Computer Vision subteam of UMARV. Most of the code was written in the 2021-2022 season.

This code contains various applications. It contains ROS nodes for lane detection, perspective transform, ROS image topic visualization, and more. It also contains the onboarding projects for the 2022/2023 season. This code was written with large help from the OpenCV python library.refer

cv_stack is maintained by the CV Subteam in ARV.

Run Commands in the Google Drive, under 2022-2023/Computer Vision will contain most commands you might want to use. This will be an explanation of the most used commands

roscore can be run before all other commands, but is usually not necessary

roslaunch zed_wrapper zed.launch will run the zed ros node with the camera plugged in. If you want to run it with an svo file, add the svo_file:="/path/to/file.svo" argument.

rosrun lane_mapping LaneDetection.py will run the lane detection ros node. You can swap LaneDetection.py with CudaDetection.py or ADSDetection.py to run those python files. Add any argument afterwards to run with pothole detection enabled.

rosrun visualize_topic visualizer.py will visualize a topic with cv2.imshow, make sure to run export DISPLAY=:0 first if you are running with ssh. If you get an error about a compressed image, use rosrun visualize_topic visualize_compressed.py instead.

rostopic list will list all of the topics currently available. rostopic hz topic_name will display a running counter of the rate that topic_name is being published.