Mite detection in ham drying rooms with AI — sweep the chamber before release to market.
Ham mites (Tyrophagus, Acarus) are invisible from a meter away, but an undetected outbreak means an entire chamber compromised. iLEAN Vision assists the drying-room lead with a vision sweep that flags hotspots before release to market. The person decides the treatment — the camera raises the alarm first.
By the time you can see the mite with the naked eye, the chamber is already compromised.
The ham mite is small, slow to detect and fast to spread. Tyrophagus putrescentiae and Acarus siro are the usual suspects, and the pattern repeats itself:
- Calendar-driven inspection — the formal round comes around every so often. Between one round and the next, the hotspot grows with nobody seeing it.
- Inspection by the operator's eye — it works, but it requires the operator to know what to look for, to get close to every piece, and to apply the same quality of review to piece number 5 and to piece number 500.
- Generic preventive treatments — applied without knowing whether there is a real hotspot, burning product and hours, and sometimes masking the problem.
- Exports to the UK and the USA — zero tolerance on sanitary appearance. A single piece with a visible hotspot can sink a shipment and damage the relationship with the importer.
The problem is not a lack of information — it is that by the time it arrives, it is already too late. And the cost of an entire chamber compromised is nowhere near the cost of early detection.
iLEAN Vision does not replace the drying-room lead — it gets the warning to them sooner.
The answer is not to replace human inspection — it is to have one more eye in the drying room, a constant one, that looks at every piece to the same standard on the first one and on the last. iLEAN acts as the putty between the formal inspection and the operator's daily routine: it fills the gap where the hotspot used to grow with nobody seeing it.
Edge sees the surface pattern of the hotspot. The agent records piece, column and height. The drying-room lead examines it and decides the treatment. The camera raises the alarm first — the person signs.
The iLEAN piece applied to mite detection in the drying room:
- iLEAN Vision (Edge) — a machine-vision terminal (CNN) trained to detect mite hotspot patterns against the real flora of your drying room (texture, granular look, fluff, uneven distribution). It mounts on a mobile cart to sweep the chamber, or fixed along the inspection aisles. It works with no network: if the drying room loses WiFi, Edge keeps sweeping and queues the trace.
- iLEAN Connect — captures the drying-room lead's voice when they confirm or dismiss a flagged hotspot (“that's dust, not a hotspot”, “this one yes, look at the granular texture”). The camera learns from the plant's real judgement, not from a generic manual.
- Agent — records every flagged hotspot with piece, column, height and timestamp. It opens the entry in the chamber's dossier, alerts the drying-room lead through whatever channel they use, and cross-references vintage, batch and commercial destination. It never triggers a treatment on its own — it proposes, the person signs.
Calendar-driven inspection vs. continuous sweeping with iLEAN Vision
| Aspect | Calendar inspection + the eye | With iLEAN Vision + Connect + Agent |
|---|---|---|
| Review frequency | By calendar; the rest, the operator's eye | Continuous, round by round, piece by piece |
| Detecting an incipient hotspot | Once it is already visible from a meter away | As soon as the surface pattern gives it away |
| Preventive treatment | Generic, just in case | Focused on the flagged pieces and zones |
| Hotspot traceability | “In chamber 3, towards the back” | Piece, column, height and timestamp |
| Risk on UK/USA shipments | Depends on the round | Last review documented before release |
| Sanitary dossier | Rebuilt by hand | Automatic, by chamber and by lot |
Impact estimate for your plant — to be validated with your numbers.
The block below is an estimate to be validated with the specific data of your drying facility. We put it forward so the committee has an order of magnitude; we refine it during the diagnostic.
- An Iberian ham drying facility with several chambers, tens of thousands of pieces, exporting to a UK or USA retailer with zero tolerance.
- Mobile Edge pilot in 1-2 chambers (terminal on a cart + integration with the ERP batch and the sanitary dossier). First value expected within a few weeks: the surface pattern starts being recorded from the very first round.
- Indicative payback between 5 and 9 months, depending on the historical frequency of incidents and the average cost of a scrapped lot or a rejected export shipment.
- Hard lever: a single shipment saved pays for the pilot. Generic preventive treatment down ≥ 30% once you move from a calendar-based protocol to a real-hotspot protocol. Conservative figures, a defensible floor.
And the drying-room lead's reasonable doubt
“What if the camera throws false positives and makes us treat when there is nothing there?” — hallucination is a problem of free generation, not of anchored tasks. When the AI simply recognizes surface patterns against a known flora and flags them for human inspection, the best models brought error below 1.5%[1]. And even so, the camera flags and the person decides: no treatment is applied on its own. The three safety rings exist precisely for this.
[1] OpenAI paper “Why Language Models Hallucinate”, 2025 — on the reliability of AI in anchored tasks.
What people ask about AI mite detection in the drying room
Which mites affect Iberian ham in the drying room, and why does it matter?
The most common are Tyrophagus putrescentiae and Acarus siro, mites under half a millimeter that colonize the surface of the ham in the drying room and the cellar. They matter because (1) an undetected outbreak can compromise an entire chamber, (2) they produce dust and “fluff” that ruin the retail appearance, and (3) exports to the UK and the USA carry zero tolerance on sanitary appearance. Early detection is the difference between one treated piece and a scrapped lot.
How does AI vision detect a mite smaller than half a millimeter?
iLEAN Vision does not look for the individual mite from a meter away — it looks for the surface pattern that gives the hotspot away: dusty texture, the characteristic granular look, incipient tunnels in the fat, uneven distribution of the fluff. The camera is trained on the real flora of your drying room (not on a generic manual) and learns to separate the normal pattern of curing from the pattern of a hotspot. When in doubt, it flags for human inspection — it does not decide alone.
At which point in the process does the sweep happen — round by round?
On every chamber inspection round, before moving the lot to the next stage and, above all, before release to market. The camera does not replace the drying-room lead — it rides along: the operator walks the round with a tablet or the Edge terminal and scans pieces as they go; the system flags suspicious hotspots for the lead to examine. Human inspection is still in charge.
Does iLEAN Vision replace accredited entomological inspection?
No. Accredited entomological inspection (external lab, microscopy, species identification) remains the formal proof. iLEAN Vision gets there earlier: it sweeps the drying room between formal inspections and alerts the drying-room lead when something does not add up, so the accredited inspection arrives with a hotspot already identified instead of searching blind. The putty fills the gap between two formal inspections — it does not replace them.
What happens when the system flags a hotspot — what is the response protocol?
The agent records the piece, the column and the height of the flagged hotspot, alerts the drying-room lead through whatever channel they use (notification, WhatsApp, phone call) and opens an entry in the chamber's dossier. The person decides the protocol: isolating the lot, an authorized treatment, calling in the accredited inspection. Nothing moves on its own: a human signature decides what to do about the hotspot.
Tell us your case and in 48h we'll send you the estimated ROI of this AI project for your drying facility.
We work on the real data from your drying facility, not ours. Diagnostic with no commitment.
Request estimated ROI in 48h See food industry