Frame ID in /tag_detections topic is empty

[tomnqduc]

I am currently working on the detection of AprilTag, however, it cannot create a transform between tag frame and the camera frame as the frame id in the message is empty string as below in the image, if someone has the same problem, I would be thankful to hear the solutions. Many thanks

For more information, I am using the video_source node from ros_deep_learning package by Nvidia to publish the image to the /video_source/image_raw topics

[Tuxliri]

I also experienced a bug with the frame_id in the detections. In my case detections[i]/pose/header/frame_id is equal to header/frame_id. The tf2 transform is correct.