Raspberry Pi Photo Booth

Eoghan0

Over the last year or so, we’ve been thinking of ways that we could use a Raspberry Pi at work, and looking for a project we could try it out with. One thought we had was a photo booth that staff and students could use to take portrait photos they could use for their VLE profiles.

Looking around the Net, I found Chris’s brilliantly detailed instructions for a photo booth he built, listed at Drumminhands.com, and decided to have a go at adapting it for our use (and also borrow heavily from the structure of his instructions).

The end result, mounted in a box created by our University joiners, looks like this:

The finished photo booth
The finished photo booth


Creative Commons Licence
Raspberry Pi Photo Booth by Eoghan Clarkson is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

The Plan

The main idea was to have an automated photo booth that staff and students could use to take profile photos. The booth would then resize the photos to the dimensions required by our various VLEs, and provide the user with a personal web adress from where the photos could be downloaded.

In addition, the photo booth could offer separate modes, such as one producing animated gifs that Chris’s photo booth provided.

So the proceedure would be:

  1. Present the user with a main menu of options, initially:
    • Official photo
      (the main mode, used for taking profile photos for VLEs)
    • Accompanied photo
      (lets users take a photo with a local landmark superimposed on it)
    • Animated photo
      (takes four photos then builds them into an animated gif, as in the Drumminhands photo booth)
    • Continuous photos
      (continuously takes photos every couple of seconds, meant as a fun activity at events)
  2. The user interface would consist of three buttons. The buttons would be illuminated to indicate when they are active, and on screen text would indicate each button’s function. On the main menu, the functions would be:
    • A button to move up the menu.
    • A button to move down the menu.
    • A button to select the current menu option.
  3. A fourth button can be used to both shutdown the Pi (single press), or quit the photo booth application (hold down for a few seconds).

The four modes would be:

Official photo

  1. An overlay appears on-screen, to show user the area their head should fill.
  2. A couple of instruction screens appear.
  3. Four photos are taken at two second intervals.
  4. All four photos are presented on screen, together with options to ‘accept’ or ‘reject’ them.
  5. If the user accepts the photos, then:
    1. They are resized into the various pre-defined sizes.
    2. They are uploaded to a web server.
    3. The web address for the images is shown on-screen.
    4. The user then presses a button to return to the main menu.
  6. If the user rejects the photos, then return to the main menu.

Accompanied photo

  1. An image appears on-screen (e.g. a statue) that the user can position themselves beside.
  2. Four photos are taken at two second intervals.
  3. All four photos are presented on screen, together with options to ‘accept’ or ‘reject’ them.
  4. If the user accepts the photos, then:
    1. They are resized into a pre-defined size.
    2. They are uploaded to a web server.
    3. The web address for the images is shown on-screen.
    4. The user then presses a button to return to the main menu.
  5. If the user rejects the photos, then return to the main menu.

Animated photo

  1. Instructions appear on-screen.
  2. Four photos are taken at two second intervals.
  3. The photos are converted into an animated gif.
  4. The animated gif is displayed on the screen.
  5. The animated gif is uploaded to the server.
  6. The user can then:
    1. Press ‘Start’ to create another animated gif.
    2. Pres ‘Exit’ to return to the main menu.

(* I should have noticed Chris’s experiences with trying to display an animated gif on screen. Instead I tried myself – but ended up just repeatedly showing each individual jpeg photo in turn, as Chris did)

Continuous photo

  1. Instructions appear on-screen, showing the web address where the photos will be available
  2. A photo is taken.
  3. The photo is uploaded to a web server.
  4. After two seconds, repeat from point 2 above.
  5. The user can press a button to leave continuous mode, and return to the main menu.

The parts

Extras

Finally

  • Create a directory on a web host, where the Raspberry Pi can upload its photos and make them available to the user. Note the FTP login details to place in the scripts.

Setting up the Raspberry Pi

I’ve put my standard set-up on the following page:

The Python code

The Code

I used Python 2.x, which was the default on the Noobs install. The code is available on my GitHub photobooth project.

The following code would copy the code into this directory:

  • ~/Documents/Python Projects/photobooth
cd ~/Documents/
mkdir "Python Projects"
cd "Python Projects"
git clone https://github.com/eoghan-c/photobooth.git

The scripts create a temporary directory on the Pi to hold the images before uploading them, in:

  • ~/photobooth

Editing from a Mac

It was much easier to use my Mac to edit the Pi’s python scripts, than to edit them on the Pi.

I connected to the Pi in two ways:

  1. Launch the Terminal application on the Mac, and connect to the Pi’s IP address, e.g.: 
    ssh pi@192.168.0.15
  2. Use the Mac’s Finder to mount the Pi’s volume:
    1. Install Netatalk, using the Pi’s Terminal command (a one-time setup): 
      sudo apt-get install netatalk
    2. From the Mac Finder’s ‘Go’ menu, select ‘Connect to Server…’
    3. Enter the Server Address, e.g.: 
      afp://192.168.0.15
    4. Click the ‘Connect’ button.
    5. Enter your Pi’s username and password

The Pi then mounts as a normal volume, and you can edit the code in your favourite editor.

Modules to Install

Before you can run the code, make sure you install the following modules:

Install the Python Imaging Library (PIL) – used for cropping and resizing images:

sudo apt-get install python-imaging

Install GraphicsMagick – used to create the animated GIFs:

sudo apt-get install graphicsmagick

Customisable parts

Photobooth_functions.py script, OfficialPhoto() function

The self.image_defs variable holds an array controlling the dimensions that ‘Official photos’ are resized to.

Each row of the array is in the form:

  • Description, image width, image height, image DPI

HTML folder

The HTML folder contains the HTML templates which are uploaded to the server along with the photos, seen when the user goes to download the photos.

Hardware

Variables in the Code

The variables defined for the hardware in config.py are (note in this project I used GPIO.BCM so the ‘GPIO PIN’ column refers to the BCM numbers):

Variable Description GPIO Pin
led_pin_select The LED of the ‘Select’ button GPIO 18
button_pin_select The close terminal of the ‘Select’ button’s microswitch GPIO 23
led_pin_left The LED of the ‘Left’ button GPIO 17
button_pin_left The close terminal of the ‘Left’ button’s microswitch GPIO 27
led_pin_right The LED of the ‘Right’ button GPIO 16
button_pin_right The close terminal of the ‘Right’ button’s microswitch GPIO 20
button_pin_exit The close ternimal of the ‘Exit’ button’s microswitch GPIO 5
The other terminals of the buttons and LEDs Ground

Fritzing Diagram

The buttons I bought contained 5V LEDs, so I didn’t need to use diodes.

Fitzing wiring diagram
Fitzing wiring diagram

You can also download the Fitzing wiring diagram.

Soldering

I worked out the wiring for the buttons and LEDs on a prototyping breadboard, together with the length of wires needed. Then soldered them into the perma-proto HAT (badly).

Perma-Proto HAT soldering
Wiring soldered into the Perma-Proto HAT board

Auto-run the Python script

I used systemd, as detailed by Matt in How To Autorun A Python Script On Boot Using systemd

(though ‘Type=simple’ seemed to work for my script).

I’ve included the Unit file with the program files, and you can use it if you have installed the scripts into the directory I used):

  1. Copy the Unit file into the correct location by using:
sudo cp "/home/pi/Documents/Python Projects/photobooth/photobooth.service" /lib/systemd/system/
  1. Then refresh the daemon with:
sudo systemctl daemon-reload
  1. And tell systemd to start the service during the boot sequence, with:
sudo systemctl enable photobooth.service
  1. Reboot the Pi:
sudo shutdown -r now
  1. And after rebooting, the Pi should automatically launch the blood_pressure scripts.

The following commands are useful to get the status, or stop the service:

sudo systemctl status photobooth.service
sudo systemctl stop photobooth.service

The above should create a Unit file at /lib/systemd/system/blood_pressure.service containing:

[Unit]
Description=A Photo Booth. Instructions at http://eoghan.me.uk/notes/2016/03/28/photo-booth/
After=multi-user.target

[Service]
Type=simple
ExecStart=/usr/bin/python "~/Documents/Python Projects/blood_pressure/main.py"
Restart=on-abort
User=root

[Install]
WantedBy=multi-user.target

Improvements

Camera Position

I put the camera module on top of the display screen – but, because the default camera module lens has a short focal length, you have to be close to the camera to fill the frame if you want a portrait photo. As the camera is placed on top of the screen, you then have to look up a bit awkwardly if you want to look into the camera.

It would be good to find a lens for the camera module that could narrow the angle of view, and allow the user to sit further way from the camera.

Lighting

I bought an LED Clip Light (£30) to stick on top of the flatscreen, that the user could switch on and off – but this doesn’t provide enough light. I think that Kevin Osborn’s idea to use LED strips as a flash might be worth a try.

Image Overlays

I displayed images on the screen using pygame.Surface with a surface alpha value – but this doesn’t allow different parts of the image to be displayed with different transparencies. So the overlay showing where to put your face in the ‘Official photo’ mode isn’t completely transparent in the centre, and darkens the whole image.

It may be better to try using per pixel alphas instead, although the documentation notes that this is the slowest drawing mode.

Also, the some image overlays are optimised for the display screen size and shape, and might not display well on a different display screen.

Customisable Variables

Move all the main customisable variables into config.py

Leave a Reply

Your email address will not be published. Required fields are marked *