Auto-Fluidics

A lightweight, open-source toolkit for real-time microfluidics experiments. It uses computer vision to track particles and includes an API for controlling various types of pumps (pressure, syringe, peristaltic). Still early days; built in spare time.

Features

• Real-time particle detection using computer vision
• Live video and image processing
• Simple API for controlling pumps
• Written in Python, with future plans for GPU/C++ support

Why?

Microfluidics gear is expensive and often locked behind proprietary systems. This project is my attempt to build something more accessible, flexible, and hackable: no special hardware or vendor lock-in required.

---

About

This repo is for anyone trying to do microfluidics without a massive budget or a PhD in lab automation.

What it does:

• Tracks particles in microfluidic setups (videos, streams, etc.)
• Analyses motion and basic metrics in real time

That’s it. It’s meant to be useful without getting in your way.

---

Project Status

Work in progress—actively developed when time allows. Right now it works with:

• Local image/video files
• Live streams (e.g., from an ESP32 camera)
• Basic CV pipelines (which are being tweaked constantly)

---

Getting Started

Requirements

• git
• Python 3.10
  • opencv-python 4.9.0.80

Clone

git clone --recursive [email protected]:finorr/auto-fluidics.git

---

Example

import cv2
import particle_detector as pdt

proc = pdt.ParticleImageProcessor()
win_name = 'Networked Stream'
cv2.namedWindow(win_name, cv2.WINDOW_NORMAL)

while cv2.waitKey(1) != 27:
    proc.process_stream(stream_ip='192.168.1.34')

    if proc.image['processed'] is not None:
        cv2.imshow(win_name, proc.image['processed'])
    else:
        cv2.imshow(win_name, proc.image['sample'])

cv2.destroyWindow(win_name)

Output

---

Build

Currently Python-only. When/if C++ support gets added (for GPU processing), this section will be updated.

---

Test

Unit tests coming soon. For now, test by running and watching for obvious errors.

---

Docs

Still being written. In the meantime, feel free to poke around or open an issue if anything’s unclear.

---

Help

If you get stuck or have questions, feel free to open a GitHub issue or contact me on LinkedIn.

---

License

MIT — see LICENSE for details.