Skip to content

Latest commit

 

History

History
142 lines (84 loc) · 4.45 KB

Readme.md

File metadata and controls

142 lines (84 loc) · 4.45 KB
Raw

Streaming video from Raspberry Pi 3(+) to AWS Kinesis Video

Camera

Install the camera module to the Raspberry Pi. Enable it in raspi-config.

amazon-kinesis-video-streams-producer-sdk-cpp

You need the kvssink plugin for gstreamer to be able to send your video stream to aws kinesis.

For this you have to build the Amazon Kinesis Video Streams Producer SDK C/C++ project.

install prerequisites

See also Install Steps for Ubuntu 17.x and Raspbian Stretch using apt-get

Install with apt-get / aptitude:

git
cmake
g++
libssl-dev
libcurl4-openssl-dev
liblog4cplus-1.1-9
liblog4cplus-dev
libgstreamer1.0-dev
libgstreamer-plugins-base1.0-dev
gstreamer1.0-plugins-base-apps
gstreamer1.0-plugins-bad
gstreamer1.0-plugins-good
gstreamer1.0-plugins-ugly
gstreamer1.0-tools
gstreamer1.0-omx

Check if your gstreamer can find the omx plugins.

pi@camera2:~ $ gst-inspect-1.0 | grep omx
omx:  omxmpeg2videodec: OpenMAX MPEG2 Video Decoder
omx:  omxmpeg4videodec: OpenMAX MPEG4 Video Decoder
...

If you don't see them. It might help to delete the directory ~/.cache/gstreamer-1.0 directory.

build

Clone the repository of the Amazon Kinesis Video Streams Producer SDK C/C++ project.

Go into the project directory, then the subdirectory kinesis_video_native_build and run:

./min-install-script

Install

When the min-install-script has run successfully. It has built a number of sample applications and a couple of libraries.

It is the latter we are interested in. More specifically:

  • libgstkvssink.so The kvssink gstreamer plugin
  • libproducer.so The producer library the libgstkvssink.so has been linked to

Create a directory to house the gstreamer plugin:

mkdir /usr/local/lib/gstreamer-1.0

Put the libgstkvssink.so in this directory. We will tell gstreamer of it's existence by adding the directory to the GST_PLUGIN_PATH environment variable.

libgstkvssink.so is linked to libproducer.so. So put the latter into /usr/local/lib. Run ldconfig -v as root. You will see that the library will be added to the ld search path.

Check

pi@camera2:~ $ gst-inspect-1.0 | grep kvssink kvssink: kvssink: KVS Sink

The streaming service

Install

Install the file from this repository by copying them to:

  • etc/systemd/system/kinesis-video.service -> /etc/systemd/system/kinesis-video.service
  • etc/default/kinesis-video -> /etc/default/kinesis-video
  • sbin/aws_kinesis_live_stream.sh

Run systemctl daemon-reload, to scan the new service.

Configure

AWS

Setup a iam user with the correct permissions. Setup a Kinesis Video Stream. See Getting Started

The raspberry pi

Edit /etc/default/kinesis-video

Set the correct access-key, secret-access-key , aws-region and stream name to use. Check that the GST_PLUGIN_PATH includes the directory where you have put the kvssink plugin.

Check / edit /etc/systemd/system/kinesis-video.service. Note the 'user'.

Create this user and add it to the video group.

Check / edit sbin/aws_kinesis_live_stream.sh. Note the LOG_CONFIG.

Copy the kvs_log_configuration from the Amazon Kinesis Video Streams Producer SDK C/C++ project's kinesis-video-native-build directory and put it in the directory where LOG_CONFIG points to.

If necessary create this directory and chown it to the user you created above.

I would recommend editing the kvs_log_configuration and modifying the root logging level. The DEBUG level produces an awful lot of logging.

Enable and start

  • systemctl enable kinesis-video
  • systemctl start kinesis-video
  • systemctl status kinesis-video

View the stream at aws kinesis dashboard

awk-kinesis dashboard

... yes it was getting dark after all this 'hacking' and writing ...

And let the fun begin

Now you can wield the power of AWS Kinesis, to do something with your video stream. Standard it will be stored with a retention of 1 day, so you can e.g. rewind to check important events captured by your camera. You can also feed the steam to other services, e.g. to recognize faces and take actions based on this.