Extract coordinates of detected object deepstream sample 1

Accelerated Computing Intelligent Video Analytics DeepStream SDK
1

hi,
How can I extract bounding boxes from detection in testing deepstream_test1_app?

is there a way to print the coordinates of the boxes on the terminal by directly changing the class c of the example?

2

In osd_sink_pad_buffer_probe function, obj_meta structure holds the coordinates for detected objects.
For example you can use (obj_meta->rect_params).top to extract the top corner of the detected object.

3

thank you so much.

I ask you another quick thing. if there are more recognized objects in the frame (2 people for example) how can I get the coordinates of both since the obj_meta object is only one? one last thing, where can I find details on these classes and functions you told me about (obj_meta etc …) thanks

[image]

4

obj_meta holds all the information for all the objects.
You can find the details here:
https://docs.nvidia.com/metropolis/deepstream/4.0/dev-guide/DeepStream_Development_Guide/baggage/group__metadata__structures.html