https://github.com/parlaynu/pi-multicapture
Record images from multiple picamera2/nvargus cameras on a central storage server.
https://github.com/parlaynu/pi-multicapture
argus-camera deepstream gstreamer jetson-nano libcamera picamera2 pyzmq raspberry-pi zeromq
Last synced: about 2 months ago
JSON representation
Record images from multiple picamera2/nvargus cameras on a central storage server.
- Host: GitHub
- URL: https://github.com/parlaynu/pi-multicapture
- Owner: parlaynu
- License: mit
- Created: 2023-11-20T03:23:06.000Z (over 1 year ago)
- Default Branch: main
- Last Pushed: 2024-01-12T03:37:22.000Z (over 1 year ago)
- Last Synced: 2025-01-28T21:34:31.392Z (3 months ago)
- Topics: argus-camera, deepstream, gstreamer, jetson-nano, libcamera, picamera2, pyzmq, raspberry-pi, zeromq
- Language: Python
- Homepage:
- Size: 17.6 KB
- Stars: 1
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# RaspberryPi Camera Streaming
Stream multiple cameras to a single server for central storage.
The system provides image streaming sources for RaspberryPi and JetsonNano and a single streaming sink which
saves jpeg images to disk. You can run a single storage sink on your network and multiple camera streaming
sources and the sink will save the images locally.
This is surprisingly simple to setup and run with the only requirements over what is needed on each system
to run their cameras in python is pyzmq.
There are a few system packages to install to get started, but I haven't started from a clean OS install on
either the raspberry pi or jetson nano so I don't have a full list. The tools will report what's missing if
you run them and it should be a simple thing to then work out what packages need installing.
I'll get around to a clean build at some point and will update the documentation here.
## Quickstart
The tools use pyzmq to send images over the network. On the pi and the jetson nano, install this with pip
in the user install location. Virtual environments won't work on these devices as they need system installed
versions of other libraries to access the camera hardware. See the `requirements.txt` file for the specific
version that I have built and tested with.
To get started, on the sink machine, run the sink server. This command below tells the server to listen for
connections on all interfaces and to write images to a directory called local.
$ ./storage-sink.py -a local
listening at tcp://192.168.1.137:8089
Then on each device with the camera, run the streamer, for example:
$ ./picam2-src.py --mode 6 --fps 2 --hflip --vflip tcp://192.168.1.137:8089
This puts the camera in mode 6 (this is 1920x1080 for my camera), captures and streams images at 2 frames per second,
flips the image horizontally and vertically so it show up the right way, and sends the images to the sink running at
address 192.168.1.137.
## Sinks
### Storage Sink
The full usage is:
$ ./storage-sink.py -h
usage: storage-sink.py [-h] [-a] [-p PORT] [-i] outdir
positional arguments:
outdir location to save images
options:
-h, --help show this help message and exit
-a, --all listen on all interfaces
-p PORT, --port PORT port to listen on
-i, --ipc use the IPC transport
This receives images from any number of streamers and writes the images to disk.
$ ./storage-sink.py -a local
listening at tcp://192.168.1.137:8089
saving local/jetson/image_0000.jpg (1359935 bytes)
saving local/pi-cam02/image_0000.jpg (633332 bytes)
saving local/pi-cam03/image_0000.jpg (577508 bytes)
saving local/jetson/image_0001.jpg (1360378 bytes)
saving local/pi-cam02/image_0001.jpg (632811 bytes)
saving local/pi-cam03/image_0001.jpg (576822 bytes)
saving local/jetson/image_0002.jpg (1361347 bytes)
saving local/pi-cam02/image_0002.jpg (633313 bytes)
saving local/pi-cam03/image_0002.jpg (577065 bytes)
saving local/jetson/image_0003.jpg (1361784 bytes)
saving local/pi-cam02/image_0003.jpg (632961 bytes)
saving local/pi-cam03/image_0003.jpg (576485 bytes)
saving local/jetson/image_0004.jpg (1361108 bytes)
saving local/pi-cam02/image_0004.jpg (632212 bytes)
saving local/pi-cam03/image_0004.jpg (632460 bytes)
### Preview Sink
Not yet implemented.
## Streamers
### Raspberry Pi Streamer
The full usage is:
$ ./picam2-src.py -h
usage: picam2-src.py [-h] [-n NAME] [-l LIMIT] [-r FPS] [-m MODE] [--hflip] [--vflip] url
positional arguments:
url the url to stream to (tcp://
: or ipc://)
options:
-h, --help show this help message and exit
-n NAME, --name NAME the name of this node
-l LIMIT, --limit LIMIT
total frames to send
-r FPS, --fps FPS camera frame rate
-m MODE, --mode MODE the camera mode
--hflip flip the image horizontally
--vflip flip the image vertically
A typical example would be:
./picam2-src.py --mode 6 --fps 2 --hflip --vflip tcp://192.168.1.137:8089
This will stream images to the sink running on host '192.168.1.137' listening on port 8089.
### Jetson Streamer
Full usage is:
$ ./nvargus-src.py -h
usage: nvargus-src.py [-h] [-n NAME] [-l LIMIT] [-r FPS] [--hflip]
[--vflip] [-c]
url
positional arguments:
url the url to stream to (tcp://
: or
ipc://)
optional arguments:
-h, --help show this help message and exit
-n NAME, --name NAME the name of this node
-l LIMIT, --limit LIMIT
limit the number of frames to send
-r FPS, --fps FPS camera frame rate
--hflip flip the image horizontally
--vflip flip the image vertically
-c, --centre crop the centre square of the image
A typical example would be:
./nvargus-src.py --fps 2 --hflip --vflip tcp://192.168.1.137:8089
This will stream images to the sink running on host '192.168.1.137' listening on port 8089.