Jam and preserves
Catch fill-level errors, label misalignment, lid seating issues, and contamination before jars leave the bottling line.
Automated quality inspection for jam, marmalade, and preserve bottling lines, running on a refurbished iPhone alongside your filler, capper, labeller, and case packer.
What is automated quality inspection for jam and preserves production?
AI defect detection for jam and preserves uses a camera and a vision model to watch every jar as it leaves the filler, the capper, the labeller, and the case packer, and to flag non-conforming units before they reach the depot. Instead of an operator at the inspection table or rigid rule-based vision, the model learns the jar shape, label artwork, fruit content, and lid geometry of your SKU portfolio, and applies a consistent visual checkpoint across shifts, line speeds, and recipe changeovers.
Jams and preserves are particularly hard to inspect at line speed because the fruit content varies inside the same batch by design, the jelly opacity reads differently across strawberry, raspberry, and apricot ranges, and the underfilled jar that ruins a multi-pack looks identical to a normal head-space variation under packing-line lighting. Rule-based vision built around a single jar shape breaks the moment you swap to a different SKU, a different label, or a different fruit recipe. 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-line sample and gives you a jar-by-jar 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 jam and preserves production lines
Fill-level errors and head-space drift
Fill-level errors are the under- and over-filled jars caused by filler-piston wear, batch-viscosity changes during the production run, or temperature drift in the jam cooker. Underfills break the case-weight specification at the retailer's depot, and overfills cause lid contamination at the capper. Operators check the filler line by eye but cannot watch every jar, so the borderline cases pass the inspection table. The AI model learns the in-spec head space for each SKU and flags drift as soon as the local fill height crosses your tolerance, with the frames available so you can adjust the filler before a full pallet ships out of spec.
Label misalignment and scuffs
Label defects include skewed application, lifted corners, glue scuffs, and creased panels caused by glue-roller wear, label-stack feed errors, or misaligned pressure rollers. The worst offenders sit on the back panel and pass the front-of-label inspection table to fail at the depot. Manual operators catch the obvious skews but miss the lifted corners that pass the labeller and fail when the wrap touches the case packer. The AI model holds the visual signature of an in-spec label for each SKU and flags skew, lift, and scuff as soon as the local pattern deviates from spec.
Lid seating and twist-off seal issues
Lid issues include cocked lids, missing safety buttons, and underseated twist-off caps caused by capper torque drift, lid-feed misalignment, or thread-roll wear. Underseated lids fail the vacuum test and lead to spoilage before the best-before date. Operators sample lids at break but miss the windows in between. The AI model learns the in-spec lid signature and flags cocked, missing-button, and underseated lids at the capper exit, with the frames available so you can adjust torque before a full batch ships.
Closure torque and tamper-band defects
Closure defects are the cousin issues of seating, including broken tamper bands, incomplete pilfer rings, and lids that spin freely on the thread. Causes include thread-roll wear, lid-batch tolerance, and capper alignment. The defects ruin the tamper-evidence claim on the label and trigger retailer rejections at the depot inspection. The AI model picks up the visual signature of a broken band or a free-spinning lid in a single frame and flags any jar that fails the spec, before the case packer wraps it.
Fruit and jelly distribution
Distribution issues include settled fruit at the bottom of the jar, jelly separation in the head space, and uneven fruit-to-jelly ratio caused by hopper agitation drift, recipe-temperature variation, or filler-nozzle mismatch. The defects ruin the consumer-shelf appearance and trigger social-media complaints. Manual operators check the filler output but miss the settled-fruit cases that pass the inspection table and look bad on the supermarket shelf two weeks later. The AI model learns the in-spec distribution for each SKU and flags drift at the filler exit so the line can adjust hopper agitation or recipe temperature.
Glass damage and chips
Glass defects include cracks at the threaded neck, chips at the rim, and inclusions in the body caused by glass-handler wear, capper impact, or supplier batch issues. The worst offenders sit at the inboard side of the case and only surface when the consumer opens the jar at home. The AI model learns the in-spec glass signature and flags cracks, chips, and inclusions at the case-packer entry so the line can divert the jar before it reaches the case.
The lighting setup that makes this work on a jam line is a diffuse overhead light over the filler and labeller to read fill level and label, plus a low-angle ring light at the capper to read lid seating. 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 jars drive a downstream divert or hold decision. We spec the optics with you during onboarding.
How Enao runs on a jam and preserves 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 lid inspection, a USB-C cable, and a mount that clamps over the filler, the capper, the labeller, or the case packer. 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 inspector on the defect families it has seen, improving with every flagged jar that the line confirms or rejects.
Each line teaches its own model what its jar shapes, label artwork, and fruit recipes look like. When you swap to a different recipe or label 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 jars stop reaching the case packer, 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 filler setup, recipe tuning, and customer complaints.
How Enao compares to manual inspection and traditional machine vision
For jam and preserves producers the comparison sharpens around five dimensions.
Setup time on a jam 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 jar and label. 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 SKUs, labels, and recipes. — 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 jars, labels, and fruit recipes in a single shift, no code to touch.
Detection accuracy on subtle fill drift and label scuffs. — Manual visual inspection: high at shift start, drops measurably after three hours. Traditional machine vision: strong on dimensional checks, weak on subtle fill-level drift and label-scuff detection. Enao: learns fill, label, and lid signatures from reference frames and holds accuracy across shifts and runs.
Who runs it. — Manual visual inspection: trained operator at the inspection table. Traditional machine vision: system integrator or a specialised vision engineer. Enao: your line team, no external specialist required.
Retailers and category managers change vendors over the cost of a rejected pallet, and the cost of a chargeback or a quiet listing swap sits well above the cost of an iPhone-based inspection rig. Enao is built for that gap.
