Cookies and Biscuits
Catch under-baked centres, broken pieces, decoration defects and bake-colour drift before trays leave the cooling tunnel.
Automated quality inspection for cookie and biscuit production, running on a refurbished iPhone alongside your tunnel oven, depositor and packaging line.
What is automated quality inspection for cookies and biscuits?
Automated quality inspection for cookies and biscuits uses a camera and an AI model to watch every tray as it leaves the depositor, the tunnel oven, and the cooling conveyor, and to flag non-conforming product before it reaches the wrapping line. Instead of relying on a tunnel-oven operator at the panel or on rigid rule-based vision, the AI learns the specific recipe, surface texture, bake colour, and topping pattern of your SKU portfolio, and applies a consistent visual checkpoint across shifts, line speeds, and recipe changeovers.
Cookies and biscuits are particularly hard to inspect at line speed because the bake-colour band sits between underdone and burnt, the geometry of a hand-style cookie varies inside the same pack by design, and the topping is meant to look generously distributed on every piece while still hitting an average count and weight. Rule-based vision built around a single recipe breaks the moment you swap to a different dough, a different topping, or a different bake profile. AI-led inspection handles those variations because the model learns from real production frames rather than from a fixed threshold.
The result is an automated visual checkpoint that complements your end-of-run QC sample and gives you a tray-by-tray image record. When a retailer query comes back six weeks later, you can pull the frames from the exact production window and either confirm the defect or push back with evidence.
Defects we catch on cookies and biscuits production lines
Under-baking and over-baking
Bake defects sit on a narrow band between an under-baked centre that fails the bite test and an over-baked rim that scorches the edge of the cookie. Causes include oven-zone temperature drift, burner imbalance, conveyor-speed instability, and dough portioning variability that changes the heat load across the band. Operators catch the extreme cases at the cooling tunnel but miss the borderline soft centre that still passes a quick glance under warehouse lighting. The AI model learns the acceptable bake-colour and surface-texture envelope for each SKU from the first half hour of a run and detects the local contrast change long before the cluster becomes obvious. Trays are flagged, the operator adjusts the oven profile, and the rejected batches get diverted before they wrap.
Bake-colour drift
Bake-colour drift is a gradual shade change across a run, caused by oven-zone aging, burner calibration drift, dough mix variability, or seasonal flour composition shifts. The first tray and the last tray of the run can sit at different LAB values without any operator noticing, and the retailer mixes packs from both windows into the same shelf set. The AI model holds a learned reference shade for each SKU and flags drift as soon as the local colour delta exceeds your spec, giving the line a chance to correct oven settings or initiate a recipe check before a tray of out-of-shade product reaches the wrapper.
Cracks and breakage
Cracks are linear surface fractures and full-piece breakage that happens during cooling, conveyor transfers, or stacking onto the wrapping line. Hard biscuits and rotary-moulded products are most vulnerable, especially when humidity at the cooling tunnel falls below spec or when a cooling-tunnel transition introduces vibration. Manual sampling at break catches the trend but misses the windows in between. The AI model picks up the surface fracture pattern on the cooling conveyor and flags the band as soon as the proportion of cracked or broken pieces crosses your acceptance threshold, so the line can divert before a wrap-rate target gets compromised on broken count alone.
Topping and decoration defects
Topping defects cover the missing chocolate chip on a chip cookie, the under-deposited chocolate drizzle on an enrobed biscuit, the off-centre cream sandwich, and the icing that bleeds outside the design line on a decorated piece. Causes include depositor dosing variability, topping hopper loading inconsistency, conveyor synchronisation issues, and humidity-driven changes to icing flow. The AI model learns the visual signature of an in-spec decorated piece from reference frames and flags the local coverage shortfall as soon as it crosses your tolerance, with the frames available so the operator can correct the depositor before half a pallet ships under-decorated.
Shape and size deviation
Shape defects cover the cookies that spread outside the design tolerance during the bake, the biscuits that come off the rotary moulder undersized after a roll wear, and the sandwich pieces that land off-centre on the bottom shell. Operators look for the obvious cases at the cooling tunnel but cannot watch every piece at line speed. The AI model learns the in-spec geometry envelope for each SKU and flags pieces that fall outside, so the line can divert before the pack fills with cookies that consumers will see as off-brand at the shelf.
Foreign material contamination
Foreign material is anything in the product stream that is not the cookie or biscuit: plastic from a torn dough scraper, a wood splinter from a flour pallet, a piece of metal too small for the detector, or a fragment of paper from an ingredient bag. Metal detectors and X-ray catch the obvious cases, but they miss low-contrast plastic and organic foreign material. A surface camera picks up the colour and texture difference against the bake background, and the AI model learns the visual signature of the materials your historical complaint record actually flags. Pieces are diverted at the cooling tunnel before they wrap, and the operator gets an early signal that the upstream process needs attention.
The lighting setup that makes this work on a cookies and biscuits line is a diffuse overhead light over the cooling tunnel to read bake colour and surface texture, plus a low-angle ring light at the topping or decoration drop to read coverage and registration. An iPhone Pro with macro and wide-angle lenses handles the seven defect families from a single inspection station per critical control point. We synchronise the rig with the conveyor encoder so that flagged batches drive a downstream divert or hold decision. We spec the optics with you during onboarding.
How Enao runs on a cookies and biscuits line
The full hardware rig costs less than €1,000 and consists of a refurbished iPhone Pro, a diffuse overhead light with an optional low-angle ring light for topping coverage, a USB-C cable, and a mount that clamps over the cooling tunnel or the decoration drop. PLC integration is not required for the first deployment, the rig fits in a flight case, and the line keeps running while you set it up.
Onboarding is self-serve. Your line team mounts the rig, opens the Enao app, and starts collecting reference frames at the next changeover. Day one returns 80% accuracy without any prior labelling, and by day fourteen the model is operating above the manual operator on the defect families it has seen, improving with every flagged batch that the line confirms or rejects.
Each line teaches its own model what its dough recipes, bake-colour palettes, and topping patterns look like. When you swap to a different SKU on the same line, the model adapts in a single shift. When you bring a sister line online with a similar product family, the second model starts from the first model's experience and the marginal effort drops sharply.
Out-of-spec batches stop reaching the wrapper, scrap is logged at the inspection point rather than at the QC office, and your operators get back the hours of attention they need for the parts of the job that still need a human, including oven setup, recipe troubleshooting, and customer complaints.
How Enao compares to manual inspection and traditional machine vision
For cookie and biscuit producers the comparison sharpens around five dimensions.
Setup time on a cookies and biscuits line. — Manual visual inspection: hours of training per operator, ongoing labour. Traditional machine vision: three to nine months of integration with a system integrator, plus a rule set per SKU. Enao: deployed in a week by your own team, day one at 80% accuracy.
Hardware cost per line. — Manual visual inspection: none upfront, ongoing labour cost. Traditional machine vision: €40,000 to €200,000 per line for industrial cameras, structured lighting, and integration. Enao: under €1,000 per line with a refurbished iPhone Pro, lamp, and mount.
Handling new doughs, recipes and toppings. — Manual visual inspection: re-train operators for every new SKU. Traditional machine vision: rewrite the rule set per recipe, often outsourced to the integrator. Enao: re-teach the model on new doughs and toppings in a single shift, no code to touch.
Detection accuracy on subtle bake colour and topping coverage. — Manual visual inspection: high at shift start, drops measurably after three hours. Traditional machine vision: strong on size sorting, weak on subtle colour drift and decoration coverage. Enao: learns bake colour and topping signatures from reference frames and holds accuracy across shifts and runs.
Who runs it. — Manual visual inspection: trained operator at the cooling tunnel. Traditional machine vision: system integrator or a specialised vision engineer. Enao: your line team, no external specialist required.
SKU rosters change with every retailer promotion and every seasonal range, and the cost of an own-label rejection or a quiet category-manager phone call sits well above the cost of an iPhone-based inspection rig. Enao is built for that gap.
Cookies and biscuits inspection FAQ
We value your privacy
We use cookies to understand how visitors use our site so we can improve it. Analytics only run if you accept. You can change your choice anytime in the footer. Privacy Policy.