mgonzs13/yolov8_ros

How can I get the image?

zhang320 opened this issue · 10 comments

How can I get the image?

Not the author but I am having the same problem as described in the title (lol)

  • ROS2 Humble running on Ubuntu 22.04, using Azure Kinect
  • nothing is being published to yolo/dbg_image or any topic really
  • launching 'ros2 launch yolov8_bringup yolov8.launch.py device:=cpu input_image_topic:=/rgb/image_formatted' and everything seems fine with one and only output '[yolov8_node-1] YOLOv8m summary (fused): 218 layers, 25886080 parameters, 0 gradients, 78.9 GFLOPs'
  • /rbg/image_formatted seems normal, I can see the camera image on RVIZ2 image viewer

Hi @zhang320 and @KXRLYM, please, check if yolov8_node is subscribed to your topic and if the QoS is the same for both, yolov8_node and your publisher.

Hello, I am using the v4l2 that comes with ROS2_ Camera sent an image_ Raw message, but I haven't received any feedback. How can I obtain tracking results?

Tracking results are published in the /yolo/tracking topic. If you don't get any results, it may be due to the QoS of your image topic. This package uses Best Effort for the image topics, the sensor QoS.

I have made the QoS consistent between my publisher and YOLO subscriber but still doesn't work :(
Added some screenshots:

No image
image
I have both my publisher and YOLO sub open

rqt_graph yolo nodes subscribing to my node
image

Publisher with

self.publisher = self.create_publisher(Image, '/rgb/image_formatted', qos_profile_sensor_data)

Thank you for your response. I have got the result.

@KXRLYM, in that case, the problem is that you are using the 3D version, which needs a depth image topic. Try the normal launch file, yolov8.launch.py.

@mgonzs13 Oh I thought I was using that launch file as per

ros2 launch yolov8_bringup yolov8.launch.py device:=cpu input_image_topic:=/rgb/image_formatted

Yes, that should launch the normal version but in the rqt_graph image, you have the 3D node. In that case, no db images will be published till 3D data is obtained.

@mgonzs13 I understood the problem now. Thank you so much the images are coming through :))