I have some iC3 and iC4 Intelbras Mibo IP cameras. Intelbras makes a mobile application available for Android and iPhone to access the cameras using its private cloud. It works, but I wanted to go further. I needed to watch my cameras on the Smart TV, without using a Digital Video Recorder (DVR) for that.
Well, I decided to run a web server and access the cameras through the TV browser. I believed it would be more useful than creating an application for Smart TV, as it could also access the cameras on my computer. The default camera protocol is Real-Time Streaming Protocol (RTSP), but I needed to use HTTP Live Streaming (HLS) to access via a web browser.
Then, I had two things to do:
- Translate, for each camera, RTSP streaming into an HLS streaming; and
- Run a web server to make the videos available.
I realized that I can make this easy with Python and FFmpeg.
Python is a versatile programming language. It has its limitations but also delivers a lot of things ready at a very low computational cost. For example, I can start a web server and run commands on the operating system in parallel, using subprocesses and multiprocessing.
To start a web server, I used the http.server
module.
As mentioned in the documentation, it is not recommended for use in a production environment, because of resource limitations that ensure better application security.
For my case, with use in the home environment, it meets.
multiprocessing
package offers concurrency, using subprocesses instead of threads. Due to this, the multiprocessing module allows the programmer to fully leverage multiple processors on a given machine.
FFmpeg is a cross-platform video and audio converter.
It has a command-line tool that makes it easy to convert between different protocols.
For example, to translate an RTSP streaming provided by an IP Camera (192.168.1.2
) into an HLS streaming, you can use this:
ffmpeg -i rtsp://192.168.1.2/live/mpeg4 playlist.m3u8
- Defines some constants, like
SERVER_PORT
,MEDIA_DIRECTORY
andRTSP_PORT
. - Gets OS environment variables:
IP_CAMERA_USER
andIP_CAMERA_PASSWORD
. - Sets the log level: info is the default.
- Cleans media directory: there, the playlists (
.m3u8
) and video segments (.ts
) of each camera will be stored. - Looks for IP cameras running in the network.
- Creates workers to start:
- an HTTP Web Server; and
- HLS Streaming cameras. For each camera:
- Creates a random ID.
- Includes camera to HTML index file.
- Starts OS subprocesses from created workers.
I coded and tested it on Ubuntu 20.04.2 LTS.
Make sure that you have Git, Python3, and FFmpeg installed. OS Environment Variables
To install them on Ubuntu, do the following:
sudo apt-get update
sudo apt-get upgrade
sudo apt-get install git
git --version
sudo apt-get install python3
python3 --version
sudo apt-get install ffmpeg
ffmpeg -version
I used:
- Git 2.25.1
- Python 3.8.5
- FFmpeg 4.2.4-1ubuntu0.1.
To make my life easier, I kept all the cameras with the same username and password.
Then, since user and password are sensitives data, they must be set as OS environment variables (IP_CAMERA_USER
and IP_CAMERA_PASSWORD
), instead of being fixed in the code.
On Ubuntu, it is possible to add in /etc/environment
file:
echo "IP_CAMERA_USER=username" | sudo tee -a /etc/environment
echo "IP_CAMERA_PASSWORD=password" | sudo tee -a /etc/environment
Restart the operating system.
git clone https://github.com/jon-cbar/eyes.git
The web server requires it to run as a superuser.
cd eyes
sudo python3 start.py
Then, open http://{your-local-ip}
with your browser to see the result.