Agriculture · Strong theory
Automated Headcount + Cast/Down-Animal Detection
Counts every animal, and shouts if one is down and cannot get up.
The physics is solid; our own validation is scheduled.
How it works
Each animal stirs the signal, so the pen count comes straight from how much the space ripples. An animal stuck on its back or collapsed goes almost completely still — the same signature as a human fall, which is already proven. Overnight, stillness in the wrong place raises the alarm within hours, not at morning rounds.
In the physics: Each animal is a moving scatterer occupying volume; occupancy count from CSI is already proven to r=0.9, and a down/motionless animal collapses its micro-Doppler to near-zero while flattening its RTI silhouette — the same physical signature as the proven human fall detector (AUC 0.8).
See it work
This is real pipeline output for Agriculture, replayed. The exact reading depends on your room.
Go deeper
How the field becomes an answer
How we prove it
Directly transfer occupancy-count (r=0.9) and fall (AUC 0.8) results from OPERAnet/SenseFi; own-hardware pen trial to retune scatterer size and count ceiling for livestock body mass.
Who it is for
'PenGuard' — overnight cast-animal and rustling/escape alarm plus free automated headcount; sells on one prevented death per season.
runtime Runs today in LatentField 2.0 as occupancy_estimate
The matching solution
Complete first-room estimateBarn Activity Node3 nodes · ~$130 parts · 90 min Request kit, subscribe, or set up →Compatible DIY hardware options
60 min · intermediate
120 min · advanced