• Hardware Platform (Jetson / GPU)
Jetson Xavier NX
• DeepStream Version
5.1
• JetPack Version (valid for Jetson only)
4.5.1
• Issue Type( questions, new requirements, bugs)
Questions
Hello Nvidia family and friends,
I’m new to Deepstream and have some newbies question.
I have a camera mounted on a gimbal, that I can control with my code.
My goal is that when the camera detects a person, the camera will follow that person.
To understand the way Deepstream works, I’m working with the python example codes from the Deepstream python apps repository.
First, I’m trying to remove the default drawing of the bounding boxes. To that end, I thought I would remove the nvosd component of the pipeline, but then the code doesn’t work anymore.
Here is the modified version of the python deepstream_test1 code (I basically commented out everything related to the nvosd object), along with the error that is displayed when I run the code.
Also, what bothers me now is that I wanted to use the metada, like the confidence of the inference, but the probe function that gets this metadata (in this case the osd_sink_pad_buffer_probe() function) is linked that nvosd object, as it is advised to do in the Deepstream documentation.
So, in order to clarify my question, how can we remove the default drawing of the bounding boxes, and how can we still get the metadata from the inference at the same time ?
Thank you very much for your esteemed help
Cheers