ROS_Deep learning (Detectnet) with deeepstream

Hi @dusty_nv, i have ran object detection module using detectnet as ROS nodes. I wanted to run object trackers and i can see that DCF and deepsort are implemented in deepstream. I wanted to know if there are any such implementation on ROS deep learning repo ([GitHub - dusty-nv/ros_deep_learning: Deep learning inference nodes for ROS / ROS2 with support for NVIDIA Jetson and TensorRT), as far as I can see, only detectors are there.

How can I use the tracker module from Deepstream into my ROS package with detector? Any suggestion other than this is appreciated to implement the tracker easily. Also, what do you think, which tracker is good to be implemented on jetson nano that takes less power as the system will be running on battery for demonstration for a short duration.

Hi @ranjeet_k, I don’t have DeepStream nodes for ROS2 myself, however you may want to check out this ros2_deepstream package to see if it meets your requirements:

Otherwise I would suggest creating the ROS2 nodes for the DeepStream components you require.

1 Like