This is a fork of the original project from the Center of HCI at the University of Salzburg.
WhiteboardBot - Whiteboard capture, enhancement and sharing tool
WhiteboardBot implements a system that uses a Raspberry Pi 3/4, or similar devices, to plug in one, or a set of cameras that would be able to capture a white- or blackboard. This capture is triggered by a BLE remote. After taking the picture, the files can automatically be sent to Slack (to a specific channel or user), to one or several email addresses or to a arbitrary mix of those.
There is also an optional enhancement feature for those pictures, such as whiteboard enhancement, lens distortion fix and cropping. It also has an optional feedback system through audio and visuals.
All those settings can be set, cameras, buttons and emails can be dynamically added and removed. Buttons can be configured, so that pressing a specific button will result in sharing through different channels. This is done through a configuration menu, which is a website that is locally reachable through whiteboardbot.local.
The guy responsible for this (mess)
Andreas Lindlbauer, formerly at the Center.
This was my bachelors project and later a project while working at the center.
There is also a bachelor thesis called "Whiteboard Documentation, Enhancement and Sharing activated by a Wireless Remote". It is about the process of building it, but may give some insight into how I found out about something or why I implemented some of the stuff, and maybe why in that way that I did. This thesis (hopefully) should be easy to get at the center. The people advising me were Verena Fuchsberger-Staufer, Martin Murer, Dorothé Smit, so they should have a copy.
I really liked working on this project so please contact me if there is something I forgot here or that you really can't figure out what to do or why I did something. Don't break my heart and just throw the code away to start from scratch without considering this project first. I made an effort to document and comment this project well.
You can reach me under these mail addresses:
How it works
For the general idea what it does and how it does that, see thesis (Whiteboard Documentation, Enhancement and Sharing activated by a Wireless Remote).
The code is also heavily commented which explains every part of the code.
Maintenance
Restart
The two cameras suck too much power on restart of the raspberry, that's why one is ignored. This is because the raspi 4 doesn't have enough amps. To fix this, just unplug the cameras and plug them in again. I know this is bad but it would need a USB 3 Hub to fix.
Access
Access is possible through two ways:
-
Plug in screen and keyboard
-
Get access through SSH while in eduroam or other university network.
- Access with:
ssh pi@<IP-address>
- The IP-address doesn't change much in the network, you can get it from the technichal assistants.
- Password is also available there.
- Access with:
Sound
Sound is controlled over the program alsamixer
.
Just type alsamixer
and press enter, which will show a menu in the terminal where you can control the volume.
The sound files are in /home/whiteboardbot/Music/sounds/
. If you want to change out the files without changing too much in the code, make sure they are .wav
wave files. If you use a different file format, you might also have to use a different program than aplay
.
Cameras
The cameras are plugged in over USB 3.0. Make sure that any cable or extension cord is compatible with that (The Lindy one isn't) and the port of any other device is also USB 3.0 else the pictures are not 4K anymore (They downgrade to 1080p).
If the pictures on the merged image are in the wrong order, either switch the USB-Plugs or look for occurrences of /dev/video0
and /dev/video1
in the code and switch them out with each other. This is the pointer to the webcam, so if anything else fails with that, it may be that one is now /dev/video2
or none at all for whatever reason.
Slack
The file slack_uploader.py
is just a wrapper for the API-Calls to the slack bot. So it is used to send messages, upload files ect. but other than that nothing complicated is done there.
The Slack company itself provides a Slack client for Python which makes it easier to connect to the bot and issue API-calls.
Log files
/var/log/whiteboardbot/bot.log
>> Saves every output of the whiteboardbot, as well as error messages, including date and timesystemctl status whiteboardbot.service
is also good to check to get more information about the system statusjournalctl -e
give you general errors of the whole raspi
Miscellaneous
- Over Ethernet it works fine, but if you change it to WiFi the IP-address will most likely change.
- You COULD give it a permanent IP-Address. The IP-Address never changed on me, so I never bothered to do that.
Sorry
- You COULD give it a permanent IP-Address. The IP-Address never changed on me, so I never bothered to do that.
Features in progress
More stable camera setup
Because the cameras are fixed on the ceiling by some flimsy wood board, the pictures aren't lined up perfectly.
My idea was to get:
- Two Metal rods from Bauhaus (Whiteboard Table)
- Gewindeadapter/screw thread adapter (Whiteboard Table)
Screw the rod to some (wooden or metal) plate, screw that plate to the ceiling. Screw that rod with the screw thread adapter to the camera. Watch out regarding the length of the table, the pipes/AC stuff and the projectors can interfere with the line of sight of the camera.
Sound
To change the volume, use the command alsamixer
over the shell, which will show a GUI inside the shell where you can change that. You may have to press F6
to change to the correct sound card aka sound device that you want to change the volume of.
Picture Quality
Except from the faulty alignment caused by the placement of the camera and settings of the distortion fixing (which also cuts the pictures as to fix alignment) there is also some glare, which could be fixed by putting on polarization filter foil on the camera.
When fixing the distortion fixing see https://www.imagemagick.org/Usage/distorts/#barrel and https://www.imagemagick.org/Usage/crop/#chop.
Fixing the barrel distortion is a little bit strange. So according to the first link you need those variables. To get these variables you have to use Hugin. This will calculate the values from putting in several pictures taken with the cameras. Don't try to guess those values, that isn't really a possibility if you value your time and sanity. Here is a tutorial, hopefully it should do trick: http://hugin.sourceforge.net/tutorials/calibration/en.shtml . They changed their GUI just as I was finishing the fix. The cameras were moved and I wanted to do that after the final setup with the rods.
The way to set the parameters for cropping the images is also a little strange, so if you want to fix alignment here, read that second link.