Raspberry Pi camera setup
The dark tinkering months have started again, and so I bought a 5 Mega pixel camera for my Raspberry Pi to use for learing OpenCV with Python. This article is a summary of how to set up the camera for wireless video streaming over your home wifi.
What you need:
- Raspberry Pi 3, other versions might work as well, please let me know in the comments
- Raspberry Pi camera v1.3, other versions might work as well, please let me know in the comments
- balenaEtcher or your referred to to flash an SD card
- Raspberry Pi OS downloaded.
Headless RPI configuration
First we’ll prepare an SD card with the help of e.g balenaEtcher to flash it with the latest minimal Raspberry Pi OS. Once the flashing is complete add and empty file named ssh to the SD card boot partition.
Also, add a file named wpa_supplicant.conf with the following content
country=<COUNTRY_CODE>
ctrl_interface=DIR=/var/run/wpa_supplicant GROUP=netdev
update_config=1
network={
ssid="<NETWORK_NAME>"
psk="<NETWORK_PASSWORD>"
scan_ssid=1
}
Be sure to edit the placeholders with the data specific to your usecase. The scan_ssd=1 line is needed when you have a hidden network.
When these two files exist and have the correct setting (please make doubly sure) save and then eject the SD card, insert it to your Raspberry Pi, and power it up
Client machine
For connecting to our server script, which runs on anather machine than the Raspberry Pi we need to gather some information. First will aquire the server ip. Run ifconfig | grep "inet " | grep -v 127.0.0.1 | cut -d\ -f2 to check your device ip (this works for my mac, go to wherever this data can be found on your machine)
Check if the Raspberry Pi has connected to your wifi by running nmap -sP <A.B.C.0>/24 where a, b and c are as in the IP you found in the previous step. When the Raspberry Pi is listed you can run ssh-keygen -R raspberrypi.local followed by ssh [email protected] this should prompt you for a password and with a correct password input, you should be loggen on the Raspberry Pi command line interface.
Passwordless SSH
It’s more user friendly and easier to securely connect over ssh by configuring keys.
Genereate a new key on the client machine with
ssh-keygen -f ~/.ssh/id_rsa_rpi
Then copy the public part to the rpi with
cat ~/.ssh/id_rsa_rpi.pub | ssh [email protected] 'mkdir -p ~/.ssh && cat >> ~/.ssh/authorized_keys'
Add the key to your keychain so it will be picked up automatically when needed.
ssh-add -K ~/.ssh/id_rsa_rpi
Log out of the ssh session if you havent already and try to login via ssh again, you should get to the cli withouh supplying a password as the authorization is now done via ssh keys.
$ ssh [email protected]
Linux raspberrypi 5.4.51-v7+
...
Camera setup
The camera I’ll setup is a cheap 5 Mpixs camera bought from Amazon. It might work for others in the same way.
While in the RPi cli, run:
sudo raspi-config
In the menu select Interfacing... -> Camera and enable it. After changing this setting the rpi needs to reboot.
After the reboot, login to the rpi again and run raspistill -v -o test.jpg to check if everythign works as expected, its output should look like this.
"raspistill" Camera App (commit f97b1af1b3e6 Tainted)
Camera Name ov5647
Width 2592, Height 1944, filename test.jpg
Using camera 0, sensor mode 0
GPS output Disabled
Quality 85, Raw no
Thumbnail enabled Yes, width 64, height 48, quality 35
Time delay 5000, Timelapse 0
Link to latest frame enabled no
Full resolution preview No
Capture method : Single capture
Preview Yes, Full screen Yes
Preview window 0,0,1024,768
Opacity 255
Sharpness 0, Contrast 0, Brightness 50
Saturation 0, ISO 0, Video Stabilisation No, Exposure compensation 0
Exposure Mode 'auto', AWB Mode 'auto', Image Effect 'none'
Flicker Avoid Mode 'off'
Metering Mode 'average', Colour Effect Enabled No with U = 128, V = 128
Rotation 0, hflip No, vflip No
ROI x 0.000000, y 0.000000, w 1.000000 h 1.000000
Camera component done
Encoder component done
Starting component connection stage
Connecting camera preview port to video render.
Connecting camera stills port to encoder input port
Opening output file test.jpg
Enabling encoder output port
Starting capture -1
Finished capture -1
Closing down
Close down completed, all components disconnected, disabled and destroyed
We can inspect the image on the client machine by running
scp -r [email protected]:/home/pi/test.jpg ~/Downloads/test.jpg
Streaming to host machine
The following example is using picamera and the code is straight from their documentation. Create two files, client.py and server.py.
The server.py script contains
# This example is based on the server code found in the picamera documentation
# https://picamera.readthedocs.io/en/release-1.13/recipes1.html#capturing-to-a-network-stream
# and has been adapted to work with opencv-python.
import io
import socket
import struct
import cv2
import numpy as np
server_socket = socket.socket()
server_socket.bind(('0.0.0.0', 8000))
server_socket.listen(0)
connection = server_socket.accept()[0].makefile('rb')
try:
while True:
image_len = struct.unpack(
'<L', connection.read(struct.calcsize('<L')))[0]
if not image_len:
break
image_stream = io.BytesIO()
image_stream.write(connection.read(image_len))
image_stream.seek(0)
frame = cv2.imdecode(np.frombuffer(image_stream.read(), np.uint8), 1)
cv2.imshow('frame', frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
finally:
connection.close()
server_socket.close()
and needs to run on the host machine where you like to view the video stream.
You should be able to run it with python server.py. It is dependend on opencv-python and numpy packages
which need to be installed before with e.g pip.
The client.py script needs to run on the RPi.
# This example is the client code found in the picamera documentation
# https://picamera.readthedocs.io/en/release-1.13/recipes1.html#capturing-to-a-network-stream
import io
import socket
import struct
import time
import picamera
# Length of stream
recording_time = 30
client_socket = socket.socket()
client_socket.connect(('your.ip.address.here', 8000))
connection = client_socket.makefile('wb')
try:
with picamera.PiCamera() as camera:
camera.resolution = (640, 480)
camera.framerate = 30
time.sleep(2)
start = time.time()
stream = io.BytesIO()
for foo in camera.capture_continuous(stream, 'jpeg',
use_video_port=True):
connection.write(struct.pack('<L', stream.tell()))
connection.flush()
stream.seek(0)
connection.write(stream.read())
# comment or remove for never ending stream.
if time.time() - start > recording_time:
break
stream.seek(0)
stream.truncate()
connection.write(struct.pack('<L', 0))
finally:
connection.close()
client_socket.close()
Replace your.ip.address.here with your actual server ip that was found earlier. Then run it with python client.py. You might get an error saying
pi@raspberrypi:~ $ python client.py
Traceback (most recent call last):
File "client.py", line 5, in <module>
import picamera
ImportError: No module named picamera
In that case install the module with sudo apt-get install python-picamera python3-picamera and try again.
When succesful, on your host(where server.py is running) you should see a screen popup with the video stream.
Cleaning up
You might have seen this message in the RPI when loggin in during experimentation:
SSH is enabled and the default password for the 'pi' user has not been changed.
This is a security risk - please login as the 'pi' user and type 'passwd' to set a new password.
To fix this add the following to the end of the /etc/ssh/sshd_config file (with e.g vim)
Match User !root
PasswordAuthentication no
Also make sure that the following lines (in the same file) are uncommented and configured with the saem values
PasswordAuthentication no
...
ChallengeResponseAuthentication no
...
UsePAM yes
So that anything that requires root access (sudo) will still ask for a password.
Conclusion
That’s all. Next steps would be to play around with opencv functionality or just use the current setup to see how plants are growing.
Comments