How ffathy-tdx/aam-cv-fsae works
Project type and purpose
• This is a small computer-vision support project for FSAE competitions. It consists of:
• Two Python data-preparation scripts (remove_class.py and count_class.py) that operate on the YOLO-formatted dataset.
• A ROS package (yolov10_cone_detection/) that runs a YOLOv10-based cone detection node in real time, with CMake build files, a ROS launch file, and a Python node implementation.
File structure
• Top level:
• LICENSE, README.md
• remove_class.py and count_class.py: standalone dataset-prep tools
• yolov10_cone_detection/: a ROS workspace package
• Inside yolov10_cone_detection/:
• CMakeLists.txt, package.xml: ROS build configuration
• launch/yolo_cone_detection.launch: launch description to start the node
• src/:
– yolo_cone_detection_node.py: subscribes to camera topics, runs YOLOv10 inference, publishes detection topics
– test.py: test or utility for the node
README for additional insights
• Dataset scripts improve label integrity and distribution before training.
• The ROS package is not documented in README, but by naming it “cone_detection” and seeing standard ROS layout, its role is inference at runtime.
Details to capture
• Launch file invoking the node and setting parameters.
• File paths (e.g. yolov10_cone_detection/src/yolo_cone_detection_node.py) as annotations if desired.