SushiRNG

Fish tank random number generator. This project uses a Raspberry Pi 3 with a camera to create a live stream of my fish tank. This live stream is then used to generate a random number.

Table of Contents

• Image Processing
  • Raspberry Pi Setup
  • Raspberry Pi Build
• Server
  • Server Setup
  • Server Build
• NIST Testing
• Mobile Application
• Configuration
• Contributions
• License

Image Processing

Raspberry Pi Setup

Install OpenCV

For this project we used a Raspberry pi 3 and installed OpenCV 3. Any Raspberry Pi and OpenCV version should work (installation took me about 6 hours). I used this tutorial from Adrian Roseblock to help me install OpenCV on the Pi.

Install Bitstream

For this project we used the python package Bitstream 2.0.3. Information about Bitstream can be located here. Please follow the instructions from that website.

Raspberry Pi Build

At this point we should have all dependencies installed on the Raspberry Pi, so go ahead and clone this repo

git clone https://github.com/FishyByte/SushiRNG.git

For the Raspberry Pi we only care about the rpi directory. Before we can fire up the OpenCV script we're going to need to tweak the color ranges in rpi/app.py. Grab some still images from the pi with the raspistill command and then use the provided range detector script to find the HSV color space range of the image. This is how the fish will be tracked.

cd rpi/testing
python range-detector -f HSV -i path/to/image.jpg

Take note of these values and adjust the color range parameters to match in the rpi/app.py file. Also while your in that file find line 97 and uncomment it out, this is turned off for production.

We are finally ready to test out the OpenCV code, make sure you are in the rpi directory and run

python app.py

Server

Server Setup

Server Build

NIST Testing

To test against our random numbers that we generate and make sure it's cryptographically secure we used the official NIST Testing Suite. (Credits to National Institute of Standards and Technology) This runs specific tests against a stream of binary bits that we have generated from our algorithm for randomness. To run these tests against your bits of data, first go the the NIST directory and run this command:

---NOTE: It would help to put your data set in the data directory---

python bitStreamTesting.py __NumberOfBitsToTest__ PATH_TO_FILE

Example command that you can try with one of our example data sets that come with this repo:

python bitStreamTesting.py 32768 data/fishBits3.txt

You can learn more about the statistical NIST Testing, the organization itself, and the NIST Test Suite at the provided link below

National Institute of Standards and Technology NIST Statistical Test Suite

Mobile Application

See our application Fish Fate. Which uses this project to generate random values.

Contributions

Open Source Community

• Adrian Roseblock

Other Sources

• National Institute of Standards and Technology
  • Andrew Rukhin, Elaine Barker, & Larry Bassham
    
    NIST DISCLAIMER

Contact Information

To contact any specific member of the FishyByte team please refer to their personal github profiles.

• Matthew O'Brien - Email: matthewo@pdx.edu
• Chris Asakawa - Email: casakawa@pdx.edu
• Nick McHale - Email: nicholas.mchale@pdx.edu
• Corey Aing - Email: caing@pdx.edu

Reporting Issues

Please report all issues on the github issue tracker

License

The MIT License