This is a door lock which can only be opened by a sonic screw driver. This uses a raspberrypi a microphone and a sonic screw driver
This project was created on 03/19/2014 and last updated 6 months ago.
Finally I don't have to worry about Time Lords being locked out of my room.
Hooked up all the components, and everything worked great. Time for the install
So the door strike arrived this week and I just wired it up and filmed a little video showing it working. Just need to get the software running on the pi and then hook everything up together....
Have you considered using the Raspberry Pi camera to grab stills and crudely detect the business end of the screwdriver's glow? "Processing" comes to mind to do this; then the Pi would detect not only the sound but also the presence of a green glow somewhere in its field of vision before unlocking.
Good call...Still haven't run the code on the Pi yet so I am uncertain how much CPU headroom i am going to have. But the sonic screwdriver I have has a green LED on the front so its likely a pretty distinct color signature to match on.
I have managed to get 4FPS pattern recognition using the picam and opencv. My starting point was here: http://thinkrpi.wordpress.com/category/opencv-and-pi-camera/ but it gets quite complicated moving around in customising opencv and userland as some point but it does work. There is another way, by using this: http://www.linux-projects.org/modules/sections/index.php?op=viewarticle&artid=14 to turn the picam into a UV4L device. Then you can use a normal opencv install on rpi to do the pattern recognition. Or you could maybe use: http://www.lavrsen.dk/foswiki/bin/view/Motion/WebHome to detection motion and then send the image to opencv to process.
That's great news! I tried and failed myself with Processing and the Pi camera. I tried the same link to turn the picam into a UV4L device but couldn't get Processing to recognize it. My backup plan of using Processing's "open" command to do a system call to run the 'raspicam' command & generate a snapshot jpg to work from failed as well and I gave up on it. Maybe Processing on the Pi (though it appears to otherwise run fine) is just a dead end.