Machine Learning on a Raspberry Pi

It looks like Google is catching up to the idea of machine learning on a Raspberry Pi! Someone put RapidMiner on a Pi back in 2013 but it was slow because the Pi was underpowered.

The Pi has been a great thin client and a small, but capable server. I’ve used it for my Personal Weather Station project and as an FTP server. Based on the news, things are about to get interesting for both Google and Raspberry Pi.

I don’t know what Google is planning to release to the Pi and Maker community but based on the survey I filled out, they haven’t decided yet. They’re looking at C#, Javascript, Go, Swift, Python, and all the other usual suspects.

The problem is optimizing the machine learning libraries for the Pi and having enough available to make it worthwhile for the community.  My guess is that they’ll go with Python, TensorFlow, and Go (Grumpy).

Whatever they decide, I consider this big news for Tinkerers and Makers everywhere. There will be an explosion of innovation if the Google toolkit is comprehensive. The Startup barrier to entry has been lowered, all you need is Pi ($40), a domain, some sweat equity, and a dream.

Raspberry Pi Project Links

Updated: 26.12.2018

This is just my running list of Raspberry Pi Projects. Commentary may be included and this lists get’s updated from time to time.

Music Related

Gaming Related

General

Raspberry Pi Projects

Pi’s are great for automation too, especially for low level stuff that doesn’t need to have an IP or be hosted on AWS. A lot of my Python scripts can easily be automated on a Pi. You can even run a server with NoIP if you want to open a port into your home network (scary). Get one today, you won’t be sorry!

Raspberry Pi Projects

Over time, I’ve amassed a small army of Raspberry Pi computers by putting them to use in all sorts of small and fun projects.

I first started out with the #RaPi project, installing and configuring RapidAnalytics on a 2 Pi cluster successfully. The experiment came to a grinding halt after I tried to initialize RapidAnalytics. The 2 Pi cluster was ridiculously underpowered and I couldn’t get it to run successfully. I theorize that that problem could be solved by adding a 3rd or even 4th Pi to the cluster and trying again.

Today, I have two Pi’s in production running my media center (Raspbmc) and another uploading my weather station (WeWxx) data to the Internet. The Pi that’s attached to my TV is quite a capable media center, and considering that we’ve cut the Cable TV cord, it’s quite a cool system. Just this past weekend I hooked up my 2TB external HD with our photos, music, and movies to it.

For my next project, I started building a game camera using a Pi. I went out and bought a camera module for a Pi that I had lying around and hooked it up. I tested out the timelapse and video capabilities and was surprised how good it is for a 5mp camera. It’s not a fancy digital camera but for what I want to do it’s plenty cool.

Using some posts I found in the Pi forums and online (see my Pinboard for them) I started cobbling together the needed scripts and services I’ll need to do this. Loosely speaking, the idea is to put a Pi with a camera module and IR motion sensor to take photos and a video when the sensor is tripped. The images and video will be stored on the Pi or transmitted via FTP if a wireless connection is available. I was thinking of using a USB thumb drive for easy removal and replacing when the system is running.

Power wise, I’m thinking of at first running the Pi on AA batteries but I have to figure out the power draw correctly. This guy’s project (@RPiAOne) really excites me because he powers his Pi completely on solar recharged batteries. Awesome!

In the end, my goal is build this game camera, test it, reproduce the system as a community project for my son’s Boy Scout den, and then donate it to a local wildlife preserve. Of course, I will share my blueprints, scripts, and Pi related stuff with you all.