Obtaining MetaData from buffer in appsink from python

[joseph]

I have a Python appsink written in Python.  I'm able to get the buffers for
each frame in the video.  It's not obvious how to obtain the MetaData that
has been generated along the way.  My pipleline looks like the following ...

gst_string = " filesrc location={} ! decodebin ! videoconvert ! gvadetect
model={} device=CPU inference-interval=10  ! gvatrack
tracking-type=short-term ! queue ! gvawatermark ! video/x-raw,format=BGR !
appsink name=appsink emit-signals=True".format(input_file, pvb_model)

I would like to be able to obtain things like the object type detected,
object bounding boxes, tracker output (object id and corresponding bounding
box).



--
Sent from: http://gstreamer-devel.966125.n4.nabble.com/
_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/gstreamer-devel

[Josh Doe]

On Fri, Jan 15, 2021 at 5:33 AM joseph <[hidden email]> wrote:

I have a Python appsink written in Python.  I'm able to get the buffers for
each frame in the video.  It's not obvious how to obtain the MetaData that
has been generated along the way.  My pipleline looks like the following ...

gst_string = " filesrc location={} ! decodebin ! videoconvert ! gvadetect
model={} device=CPU inference-interval=10  ! gvatrack
tracking-type=short-term ! queue ! gvawatermark ! video/x-raw,format=BGR !
appsink name=appsink emit-signals=True".format(input_file, pvb_model)

I would like to be able to obtain things like the object type detected,
object bounding boxes, tracker output (object id and corresponding bounding
box).

Wouldn't you just use the Python bindings here? Haven't tried GVA myself, but I should look into it.

https://github.com/openvinotoolkit/dlstreamer_gst/wiki/Python

 

_______________________________________________
gstreamer-devel mailing list
[hidden email]
https://lists.freedesktop.org/mailman/listinfo/gstreamer-devel