Inspiration
We had an Intel Edison machine, so we wanted to use that and do a hardware hack. We borrowed a webcam from Intel, and got a lot of useful advice from them. We also wanted to use Wolfram's image/face recognition and language.
How it works
We have a webcam hooked up to an Intel Edison. It uses ffmpeg to capture a photo, which it then uploads it to Amazon Web Services using Node.js. We have a website that retrieves these images from Amazon S3 and displays them.
Challenges
The hardware kept breaking. The Edison came with a stripped version of Linux, so we had to install several packages to have a functional operating system. Even after that, we had to install things sort of manually because we did not have apt-get, yum, or pip functionality. We used edi-cam and ffmpeg to get the webcam to communicate with the Edison and could live stream after the first night, but the webcam stopped working the net day because of wifi issues. At one point, the Edison's kernel modules broke and we had to reflash the Edison and reinstall everything.
Accomplishments
We managed to get the hardware to work even after having to work backwards for several hours. Because of hardware troubles, we did not have enough time to implement our original ideas of motion sensing for music or facial recogniton. However, we managed to create a somewhat successful hack on hardware and software we had never used before.
What we learned
The Edison is an external Linux machine, so we had to use puTTY and Linux to interact with it. We had to utilize many of the functions of the Linux system, including shell scripts, Vim, and using it as a server. We also learned about the hardware itself. We tried out several services and APIs that were confusing but very powerful.
What's next
We have a website set up that uploads all the pictures taken with the EdisonCam. It can be a cheap way to take and upload photos from a set location without needing an actual computer and camera.
Log in or sign up for Devpost to join the conversation.