Food factories

Computer vision for food factories

aisabela connects to the cameras the factory already has and alerts the line supervisor, quality control or the warehouse when something goes out of spec: portion weight, baking, defects or supplier receiving. Production-line quality control for food manufacturing, shift after shift.

Updated:

Sorting lines in a food factory, seen from the camera
Factories. Lines 1 and 2 are within weight; on line 3 pieces weigh 112 g against a 100 g ±5 g target. Alert to the line supervisor: adjust the line 3 depositor.

Everyday problems in a food factory

Quality, portion weight and compliance on the line, shift after shift.

  • Off-weight pieces

    Found after packaging, when the whole batch is out of range.

  • Uneven baking

    Color and bake change with the oven, the shift or the time of day.

  • Defects that reach packaging

    A broken or misshapen piece moves on because nobody saw it in time.

  • Every shift, its own method

    The recipe and process are applied differently on each shift.

  • Unchecked deliveries

    Packages are missing or goods arrive in poor condition, and it is noticed too late.

A day of alerts in a factory

This is a normal day with aisabela: the line supervisor, quality control, maintenance or the warehouse get the task, resolve it and it is logged.

One day of alerts in a food factory (example)
TimeAlertSent toStatus
06:30 Morning shift: full PPE Line 2 OK
08:14 Line 3 off-weight (112 g) Alert to the line supervisor Resolved
10:51 Defective piece before packaging Alert to quality control Removed
13:02 6-minute stop at the packing machine Alert to maintenance Resolved
15:20 Receiving: 2 packages short of the delivery note Alert to the warehouse Reviewed

Illustrative example: times, locations and figures do not belong to a real customer.

Which areas of the factory does it cover?

From receiving to packaging, with the cameras you already have.

  • Forklift between warehouse racks

    Receiving

    Packages and quality checked against the delivery note.

    Receiving: 2 packages short of the delivery note Alert to the warehouse

  • Loaves on a production tray

    Production line

    Weight, size and defects of every piece.

    Line 3 off-weight (112 g) Alert to the line supervisor

  • Pastries inside the oven

    Ovens

    Color and bake within the standard.

    Baking outside the standard Alert to the line supervisor · batch 14

  • Freshly baked loaves on a rack

    Packaging

    Defective pieces removed before packing.

    Defective piece before packaging Alert to quality control

  • Warehouse aisle with tall racks

    Raw material storage

    Stock levels without manual counts.

    Flour below minimum in the storeroom Alert to purchasing

Who gets each alert in the factory?

Everyone sees what is theirs to handle, from the line to plant management.

  • Line supervisor

    Gets the alerts for their line—weight, defects, stoppages—and adjusts on the spot.

  • Quality control

    Removes defective pieces and has a record of every deviation by batch, line and shift.

  • Plant management

    Every line and shift on one screen, with data on what happened, where and when.

Is there a real factory case study?

Yes: Malvón Factory, the Empanadas Malvón factory.

At the Malvón factory, aisabela checks every empanada for a consistent size. It is the same piece weight and size check that can be applied to other product lines.

  • 1 by 1 every empanada checked for a consistent size

Source: Malvón case study.

Computer vision in food manufacturing FAQ

Does it replace the quality control I already have?

No, it strengthens it. Your sample checks stay as they are; aisabela watches the line continuously and alerts quality control when a piece goes out of spec, with the image and the batch.

Does it work with any product?

It works with what is visible in the image: size, shape, bake color, visible defects or packages at receiving. When we review your line, we tell you what we can check with your cameras.

Do we have to stop the line to install it?

No. aisabela connects remotely to the cameras you already have, with no construction work and no line stoppages.

What about line workers’ privacy?

aisabela uses no facial recognition and works with 100% non-biometric data: it looks at the product, the line and the process, not at who each person is.

Who are you?

Try it with your own cameras

Tell us about your business and we will show you, on your own cameras, the issues aisabela would catch in your day-to-day.

Add aisabela’s AI to your projects

Your clients get more out of their cameras and you open a new line of business. For installers, integrators and distributors.

Controller: ISABELA LABS SL (company in formation). Purpose: to handle your demo, contact or partnership request. Legal basis: consent. Recipients: no data is disclosed to third parties unless required by law. Your data is processed by our technology provider Microsoft, with an international transfer to the US covered by the EU-US Data Privacy Framework. More information in the privacy policy. Rights: access, rectification, erasure and others, see the privacy policy.