How to show result detection (yolov3) on Rviz [closed]

asked 2019-10-16 06:35:09 -0600

tiensu gravatar image

Hi everyone!

I want to test Yolo detection on Autoware 1.12.0 with LGSVL simulation 2019.09.

I saw the image data is published on /simulator/camera_node/image/compressed topic from Simulator.

On darknet_yolo3 window, I config:

  • image_src: /simulator/camera_node/image

  • cfg: /home/autoware/Autoware/install/vision_darknet_detect/share/vision_darknet_detect/darknet/cfg/yolov3.cfg

  • weights: /home/autoware/Autoware/install/vision_darknet_detect/share/vision_darknet_detect/darknet/data/yolov3.weights

  • name: /home/autoware/Autoware/install/vision_darknet_detect/share/vision_darknet_detect/darknet/cfg/coco.names

vision_darknet_detect node runs successfully.

On ImageViewerPlugin window, I chose the following topic:

  • Image Topic: /image_raw
  • Object Rect Topic: /detection/image_detector/objects

The image from Simulator is displayed, but I can't see result detection (bounding box).

Can you show me what I'm wrong?

Thank you very much!

edit retag flag offensive reopen merge delete

Closed for the following reason the question is answered, right answer was accepted by tiensu
close date 2019-10-18 03:28:38.999608


Are you using some custom setup? If no can you point me to what you are using (github repo). Is YOLO subscribing to the image topic coming from simulator?

Choco93 gravatar imageChoco93 ( 2019-10-16 08:09:19 -0600 )edit

Hi, I used Autoware docker prebuilt 1.12.0:

Is YOLO subscribing to the image topic coming from simulator?

Yes, YOLO subscribes to the image topic coming from the simulator?

tiensu gravatar imagetiensu ( 2019-10-16 11:43:24 -0600 )edit

Surprise, today I can show detection results on ImageViewerPlugin with the same config as yesterday. I don't know why. Anw, thank for your attention! :)

tiensu gravatar imagetiensu ( 2019-10-16 22:51:41 -0600 )edit