We made our first test, without audio. Purpose of it is to test and calibrate software.
In general it works well, but there are few problems which cannot be solved. We are going to do everything possible to make instalation as perfect as posible.
Firs problem is that camera has angle and it recognizes people like white spots on dark matter, it does affect interaction, because people who are standing behind each other is recognized as one spot.
Second problem is that camera is very sensitive for light, it mean that if lighting changes, camera settings have to be reconfigurated.
Also it recognize shadows as white spots and can connect multiple people in one white spot.
From all of this: Instalation can work during daytime, but could glitch from time to time, best expierience could be at the night or when mainly cloudy.
For our physical part of the installation we are going to need a wireless connection between the button (which will be situated outside) and the laptop (which will be situated inside the Wirelab quarters). The 1st thought that we had was to use a Raspberry Pi for this functionality, but after a technical discussion with the teachers we decided that a wireless doorbell and a connection with arduino will do the job way faster and less complicated.
So, I bought a wireless doorbell from a supermarket and I’ve started disassembling it.
The doorbell is made out of 2 parts: the button and the receiver. On the button part we’ve connected an arcade button for a longer use. On the receiver part I’ve connected a led to visualize the tests and ,of course, the Arduino. The speaker was removed thus not being of any use.
Through the Arduino I’ve tested what voltage does it receive after pressing the button. There are voltage fluctuation thanks to the connected button that supposedly is changing the song of the doorbell.