Part Number:AM5728
Hello Guys,
Actually i have used opencv application on AM5728 evm board and stream rtsp source.But it is too slow.
So, i decide to use hardware codec to convert video format. So, i have used VPE with gstreamer and testing.
Now, if i use "videoconvert" in gstreamer pipeline with filesink and pass that file to opencv. So, its work proper.
But if i use "vpe" instead of "videoconvert", then that file could not open by opencv.
I have used "command" as below:
1).gst-launch-1.0 -v rtspsrc location=rtsp://888888:888888@192.168.1.2:554 ! rtph264depay ! h264parse ! ducatih264dec ! videoconvert ! 'video/x-raw, format=(string)BGR, width=(int)640, height=(int)480' ! filesink location=video_test.bgr -v
2).gst-launch-1.0 -v rtspsrc location=rtsp://888888:888888@192.168.1.2:554 ! rtph264depay ! h264parse ! ducatih264dec ! vpe ! 'video/x-raw, format=(string)BGR, width=(int)640, height=(int)480' ! filesink location=video_test.bgr -v
Now, i am attaching 2 application which i have using.
In "g_opencv.cpp", i have using gstreamer pipeline and passing frame to the opencv.
In "objectdetection.cpp" application, i am trying to pass raw-data file which has been generated by gstreamer+vpe+filesink pipeline.
So, is there any other option which should i use or how can we solve this issue?.
I have also found that "facedetect" can be use with gstreamer to detect face, but when i execute "gst-inspect-1.0 facedetect", i got log as below:
root@am57xx-evm:~# gst-inspect-1.0 facedetect
No such element or plugin 'facedetect'
root@am57xx-evm:~#
Regards,
Kishan Patel.(Please visit the site to view this file)(Please visit the site to view this file)
