Explore a distributed computer vision pipeline built using Python, OpenCV, and ZMQ for managing a small permaculture farm in Southern California. Learn how 20 Raspberry Pi computers with PiCameras and sensors capture images, detect motion, and select relevant data for tasks like reading water meters, optimizing water use, and tracking wildlife. Discover the hub and spoke design that distributes processing across devices, the advantages and disadvantages of PyZMQ messaging for image transfer, and useful OpenCV techniques for both Raspberry Pi and hub computer processing. Gain insights into hardware setups, including light types and enclosures, and access the open-source project on GitHub for code, hardware instructions, and farm photos.
Yin Yang Ranch - Building a Distributed Computer Vision Pipeline Using Python
Overview
Syllabus
Intro
Yin Yang Ranch: A Distributed Computer Vision System
Building a Small Permaculture Farm in Suburbia
Jeff Bass Bio Bullets (College, Career, Retirement)
How Computer Vision helps manage my small permaculture farm
What is a Distributed Computer Vision Pipeline?
My Tools for Computer Vision in Python
Image Computing Toolset: OpenCV
imagenode image ZMQ imagehub Pseudocode for imagenode
DCVP Example: Reading my Water Meter
Imagenode Code Snippet (Runs on a RPI)
There are many imagenode CV settings... Parameter Tuning to optimize detection
For each image captured, imagenode: • Applies Transformations
Image Transport: ZMQ
Imagehub Code Snippet (Runs on a Mac)
DCVP end to end example: the Barn Coyote Cam
Communicating to End Users (me :-)
Yin Yang Ranch DCVP: Lessons Learned So Far
Hardware: Example Light Types
Hardware Enclosures: Use Existing Light Fixtures!
Learnings: Some Hardware Tips and Tricks
Yin Yang Ranch DCVP projects are Open Source & on GitHub
Questions about this talk? Get Answers!
Taught by
PyCon US
