ROS Resources: Documentation | Support | Discussion Forum | Service Status | Q&A answers.ros.org

detection package

Hi! I have been reading the paper lately named “Autoware on Board: Enabling Autonomous Vehicles with Embedded Systems”. Here is a graph which is Figure 5. I have a question about part of the route in the detection part of the graph, which is camera -->image -->image detector --> image objects. I am wondering what packages under the detection package of Autoware are related to node named image detector and topics named image and image objects. Thank you very much.

I’ve not read the paper yet, but I guess vision_detector:
https://github.com/CPFL/Autoware/tree/a4b6f377a12313d6fd3d92ee70a5d25ac20ef8b5/ros/src/computing/perception/detection/vision_detector/packages

is the package.

/image_raw and /detection/image_detector/objects are the topics you mentioned.

https://github.com/CPFL/Autoware/blob/a4b6f377a12313d6fd3d92ee70a5d25ac20ef8b5/ros/src/computing/perception/detection/vision_detector/packages/vision_darknet_detect/src/vision_darknet_detect.cpp#L354

Thank you for your response. I will look at the references you provided to me. Also, I have a short question about this. The question is below.

Question: Are you meaning that the tasks from image–>image detector–>image objects happen inside Autoware? In other words, Are they done by Autoware rather than ROS?

@BRIGHT_BULD That’s right