Avionics meerkat Tk1 board and nvidia tools and libraries

Autonomous Machines Jetson & Embedded Systems Jetson TK1
1

I am using the avionics meerkat TK1 board and have installed their L4T version which can be found here: GitHub - avionic-design/linux-l4t: The l4t linux kernel

However, this does not have any of the CUDA tools or the nvidia libraries (I am particularly interested in the gstreamer plugins for H.264 encoding/decoding). The documentation I found seems to be about installing everything through jetpack on a ubuntu system. However, I need to use this custom kernel from Avionics. Is there any way to get the gstreamer and cuda tools on it without having to replace the OS as well. I have been going silly over this, so any help would be greatly appreciated.

2

Hi,

Thanks for your question.

You can install CUDA toolkit only via Jetpack custom actions.

For gstreamer, here is document for your reference:
http://developer.download.nvidia.com/embedded/L4T/r21_Release_v5.0/Docs/L4T_Jetson_TK1_Multimedia_User_Guide.pdf?autho=1491539927_1f644b8a9d28b3af49c6b0c799844a1c&file=L4T_Jetson_TK1_Multimedia_User_Guide.pdf

On our official image, gstreamer can be installed by following command. You can check if it also works on your environment.

$ sudo add-apt-repository universe
$ sudo add-apt-repository multiverse
$ sudo apt-get update
$ sudo apt-get install gstreamer-tools gstreamer0.10-alsa gstreamer0.10-plugins-base  gstreamer0.10-plugins-good gstreamer0.10-plugins-bad gstreamer0.10-plugins-ugly
3

Thank you for your answer. Using this GStreamer command will install the nvidia jetson TK1 specific encoder/decoder plugins? out of curiosity, is there any other way to use the HW accelerated encoder/decoder other than gstreamer?

Many thanks again for your support.

4

Yes, for example, accelerated H.264 decode can be launched with command:

$ gst-launch-0.10 filesrc location=<filename.mp4> ! qtdemux name=demux demux.video_00 ! queue ! nv_omx_h264dec ! nv_omx_hdmi_videosink -e

Please find more details in the document.

Another options is to use MMAPI, which with V4L2 for low-level encode/decode.
Sample is also available on JetPack.