This is yet another DIY project for a self-balancing robot that can track objects autonomously, classify the objects via Tensorflow, Inception V3 model, and store the resulting metadata in a blockchain.
A youtube video of the robot can be found at: https://www.youtube.com/watch?v=fvr28leoJzE
The concept for this project was to show how a blockchain could be used when an organization wanted to improve the quality of their image classifications. In this demo the business-network is made up of the robot, Jetson TX2 and a raspberry pi in the role of an analyst.
An overview of the project
Below is the schema for storing image metadata in the blockchain with the participants are mapped to which attributes they create and or update
These sites were very useful and provided information that allowed me to build DIY3: https://www.pyimagesearch.com/2015/12/14/installing-opencv-on-your-raspberry-pi-zero/ http://www.brokking.net/yabr_main.html https://www.youtube.com/watch?v=D7lkth34rgM https://www.youtube.com/watch?v=KxZ-e6G7INg https://hyperledger.github.io/ https://www.freecadweb.org/wiki/Category:Tutorials https://www.prusa3d.com https://ultimaker.com/en/products/ultimaker-cura-software https://medium.com/@julsimon/keras-shoot-out-part-2-a-deeper-look-at-memory-usage-8a2dd997de81 https://github.com/tshino/compare https://websockets.readthedocs.io/en/stable/intro.html#python-lt-36 https://www.thingiverse.com/thing:2829553 https://tutorials-raspberrypi.com/raspberry-pi-ultrasonic-sensor-hc-sr04/