search

dusty-nv / ros_deep_learning

Deep learning inference nodes for ROS / ROS2 with support for NVIDIA Jetson and TensorRT
887 stars 258 forks source link

/detectnet/detections/results/id never goes back to '0' (background), instead stops publishing #55

Closed WaldoPepper closed 3 years ago

WaldoPepper commented 3 years ago

I use this node (trying ;-) to steer a R/C car (basically a JetRacer) thru a circuit and it looks promising.

The only issue I have at this point is, that if the car loses track of the circuit and none of the trained markers is visible anymore, my code expects the /detectnet/detections/results/id going back to 0 (which is background by default) and so triggers some sort of recovery maneuver.

Instead the node stops publishing and so "keeps" the last identified object on the topic.

Is there a solution to signal "I don't see anything I was trained to see"? I could use a timeout to catch the absence of messages on the topic, but maybe there is something more elegant... ;-)

WaldoPepper commented 3 years ago

Thinking hard about it, the actual behaviour is better to be kept this way ;-)