MyceliumVision — contamination detection for mushroom cultivation, by DNTD Dynamics.
In development
What it is
Cultivating mushrooms at scale means catching contamination early — mold, bacterial blotch, competitor fungi — before it spreads through a grow. MyceliumVision is a computer-vision pipeline that watches cultivation jars and flags contamination visually, before it’s obvious to the eye.
Lion’s Mane (Hericium erinaceus) is the primary subject, grown and monitored on the bench at DNTD.
Current stack
- PyTorch + OpenCV for image classification
- Running on the NucBox K6
- SQLite for observation logging
Where this connects
Contamination events logged here are the motivating data for DNTD’s genetics research thread — the biological observations MyceliumVision generates are what the PySCF quantum chemistry work eventually investigates at the molecular level (hericenones and erinacines, the neurogenic compounds Lion’s Mane is known for).
Status
Early-stage. No dataset, model, or field results published yet — this page will grow into build logs and results as the work gets there, the same way the ForeForce kit’s documentation did.