Menu
Book a demo
Computer Vision for Egg Packing: Where an OEE Camera Fits

Computer Vision for Egg Packing: Where an OEE Camera Fits

Crack and dirt detection belong to your grader; see where an OEE camera counts packs and cases, gives video of line stops, and what to ask each vendor.
Computer Vision for Egg Packing: Where an OEE Camera Fits

In short: computer vision for egg packing means two separate systems. The grader's own inspection modules find cracks, dirt, leakers and blood, while an OEE camera counts packed output and gives you the video for the minute of a stop at the packers, the hand-pack lane and the palletizer.

  • Crack, dirt, leaker and blood detection are inspection options sold by grader makers such as Moba and Sanovo. Fabrico does not do them.
  • The grader already counts eggs at full speed. Measure OEE there, and put cameras downstream, where packs and cases leave the lanes.
  • The best first camera is usually over the hand-pack lane, which nothing counts, or the case and pallet area, where stops get blamed on the grader.
  • Fabrico builds OEE from machine signals and the stop reasons operators enter. Its camera add-on counts output; where it covers a station, the video for the minute of a stop can be opened from the shift timeline, streamed on request and not stored in Fabrico.

An egg grader is already full of sensors. It weighs every egg, and with the right options it checks each one for cracks, dirt, leakers and blood before the egg reaches a packing lane.

So when someone offers an egg packing centre "computer vision", ask which job they mean. Inspecting the egg is the grader maker's work. Counting what gets packed, and showing why the line stopped, is a different job for a different camera.

For the OEE rules first, such as grader-level OEE in eggs, size mix and downgrades, read OEE for egg grading and packing. This page covers cameras.

Egg inspection and OEE cameras: who does which job

Grader makers sell inspection as options on the grader. The spec sheet for Moba's Omnia PX lists an Egg Inspector for leakers and dirt, a Shell Inspector for cracks and shell strength, and a Blood Detector. Moba also says it now offers AI in vision weighing, egg inspection and crack detection on two of its larger models, in a PX+ version it says is only available in a limited region.

Methods differ from maker to maker. Sanovo notes that most crack detection systems today rely on an acoustic measurement, and sells a vision-based crack and leak module, the VisionAI Detector, rated at 288,000 eggs per hour.

Either way, these modules are part of the grader. The grader maker sets them up, tunes them and supports them. Fabrico does not inspect eggs, and it does not change what the grader decides.

JobGrader's inspectionOEE camera
Weighing and size gradingYes, its main jobNo
Cracks and shell strengthYes, crack detection optionNo
Dirt and leakersYes, inspection optionNo
BloodYes, blood detection optionNo
Counting packs at a hand-pack laneNot its jobYes
Showing why a packer or palletizer stoppedNoVideo for the minute of a stop
Counting cases on the belt to a hand-built palletNoYes, where nothing else counts

The two systems share one number: the downgrade count. If the grader gives a count of eggs taken out, OEE software can use it, but only as a number. The grader's crack detector can't tell you whether an egg cracked in the hen house or on your loader. That split is a rule you set, and the egg OEE guide covers it.

Track OEE and downtime in real time
Book a demo

Where an egg packing centre loses time

Egg packing runs against the calendar. Under the EU egg marketing rules, Commission Delegated Regulation (EU) 2023/2465, Class A eggs must be graded, marked and packed within 10 days of laying, as Ireland's food safety authority, FSAI, sets out. Eggs that miss it become Class B, which go to industry, not to shoppers.

The flow never stops either. The European Commission counts more than 350 million laying hens in the EU, producing close to 6.7 million tonnes of eggs a year. Hens lay every day, so a lost hour at the grader becomes eggs waiting in the cooler, and the oldest of them closer to the 10-day limit.

At grader speed, a few minutes is a lot of eggs. Moba lists its Omnia PX graders from 45,000 to 255,000 eggs an hour. At 255,000 an hour, a five minute stop leaves 21,250 eggs waiting.

Infeed: farm belts and tray loaders

Eggs arrive on belts from in-line houses, or on trays and pallets from other farms. A loader or destacker feeds the grader. A jam here starves the whole grader, and the grader's signal shows only that eggs stopped arriving.

The grader and its lanes

When a lane backs up, the grader can't drop eggs there. On the Omnia, for example, Moba says the extra eggs are diverted or stay in the buffer for one more revolution. What your grader does when several lanes are full at once is a question for its maker.

That is why lane stops belong on the same timeline as the grader. If the grader controller shares lane status, it tells you which lane was full. Video for that minute tells you why: a carton jam, a lid that did not close, or an empty magazine.

Packers and customer changes

Each lane ends in a packer that fills and closes cartons or trays, often with a printer or labeller for the customer's pack. Retail, food service and own-brand packs can all run on one grader, sometimes on neighbouring lanes.

A customer change on one lane is a stop for that lane, and maybe a slowdown for the grader. Give it its own reason and a standard time, so overruns are visible.

The hand-pack lane

If your centre has a lane or table for outsize eggs, specials, re-packs or trays for bakeries, it is probably the one station nothing counts. People pack there by hand, and the count is a tally or a guess from the pallet.

That is the clearest case for an OEE camera in an egg packing centre. It counts packs or trays at the lane outfeed, so the lane's output joins the line's numbers.

Cases, stackers and pallets

Packed cartons go into cases or onto trays, then a stacker and a palletizer build the load. When the palletizer stops long enough to fill the buffers, cases back up, then the packers, then the lanes, then the grader.

The grader's signal sees its own stop, not the cause three machines away. For where cameras go at that end, see case packer downtime on camera and palletizer downtime on camera.

Line map: from farm belt to pallet

Here is a typical egg packing centre, with the system that owns each job. The dashed box is the grader and everything built into it.

An egg packing centre, job by job Farm belt to pallet, top to bottom Signal Camera Inspection Farm belt, loader Signal Camera Grader, weighing Signal Crack, dirt, blood Inspection Packing lanes Signal Camera Hand-pack lane Camera Case packer Signal Camera Stacker Signal Camera Palletizer, wrapper Signal Camera Dashed box: the grader maker's machine, with its inspection modules.
Green: a machine signal counts and times the stops. Blue: the OEE camera add-on, for a count or for video. Amber: the grader's own inspection. The dashed box is the grader maker's territory.

Read the map in three zones. The grader and its inspection modules belong to the grader maker. The loader, packers, case packer, stacker and palletizer have signals that time their stops. The hand-pack lane has nothing, and that is where a camera count does the most.

Camera, signal or operator at each egg packing station

A first pass for each grader line. For more detail by machine type, see camera or sensor by machine.

StationCounts and stops fromWhere a camera helps
Farm belt, loaderLoader signalVideo for loader jams
GraderGrader run signal and egg countNot for counting
Crack, dirt, blood checksThe grader's inspection modulesNone. Downgrade counts feed OEE
Carton and tray packersPacker signal per laneVideo when the fault code is vague
Hand-pack laneNothing, or a tallyCounts packs or trays
Case packer, stackerPLC run stateVideo to tell own stops from blocked
Palletizer, wrapperPLC run stateVideo for the first cause

In every row, the reason for a stop comes from a person. The signal says when and for how long. The operator says why, with the video to check.

Want to see it? Book a demo.

Worked example: one shift at a 90,000 egg an hour grader

This is sample data for one shift at a grader rated at 90,000 eggs an hour, which is 1,500 a minute. It describes no real centre, and it is not a product screen.

One shift: where the stops started Illustrative, sample data Palletizer: 18 min Case packer: 14 min Grader stop for pack change: 12 min Lanes 3 and 4 full, jam: 10 min Loader jam: 6 min After the lanes: 32 of 60 stop min The grader saw these as its own stops OEE: 86.7% x 92.0% x 99.4% = 79.3% Ideal: 1,500 eggs a minute 32 min = 48,000 eggs, about 133 cases
Illustrative, sample data. Red bars started after the packing lanes, amber a grader changeover, light green full lanes, white the infeed. Bars share one scale.

Planned time was 450 minutes, and the grader's signal timed 60 minutes of stops. Availability is 390 of 450 minutes, 86.7%. Performance is 538,200 eggs counted of 585,000 ideal, 92.0%. Eggs the line itself broke came to 3,200, so quality is 535,000 of 538,200, 99.4%.

OEE is 79.3%, the same as 535,000 good eggs of the 675,000 possible. Eggs that arrived cracked or dirty stay out of quality and are reported by flock, as the egg OEE guide sets out.

Now the camera question. Of the 60 stop minutes, 32 started after the lanes: 18 at the palletizer and 14 at the case packer. The grader saw them as its own stops. If only the grader is connected, those minutes get booked as "grader" or "lanes full". With the case packer and palletizer signals on the same timeline, you see where a stop started, and the video shows why, such as a case jam or a late pallet.

At 1,500 eggs a minute, 32 minutes is 48,000 eggs, or about 133 cases of 360 eggs. The eggs still need packing, on overtime or on the next shift. That is capacity the centre paid for and did not use, and it is the number to show whoever owns the end of line.

Infographic tracing an egg line stop back from the palletizer: cases back up, packers stop, lanes fill, the grader stops. A grader signal alone shows a grader stop; signals on each machine show where it started; camera video shows why.
A palletizer stop reaches the grader last. Signals on each machine show where it started; the video shows why.

Cameras in an egg room: dust, wash-down and the grader's optics

Egg rooms are dusty, and graders get washed down. An OEE camera needs a housing that survives both, and a lens someone checks on a schedule.

Mount it over the packers, the hand-pack lane or the pallet area. Keep it, and any extra light, well away from the grader's inspection modules, so it can't disturb the grader maker's own cameras and lighting. Ask the grader maker before you fix anything on or near the grader.

A sensible start is one camera over the hand-pack lane and one over the pallet area. Plan the install around your biosecurity rules, like any contractor visit. The guide to how many cameras a line needs helps size the rest.

Questions to ask any vendor of computer vision for egg packing

Ask every vendor the same questions. Bring a week of grader reports and one shift's tally from the hand-pack lane.

AskHow to test the answer
What does it need from the grader?Ask what it reads from the grader controller, and how it would count eggs and time stops if the grader maker won't share data.
Which count is the line's count?Expect eggs at the grader. Ask how packs and cases turn back into eggs for each pack format.
Can it count the hand-pack lane?Film a run of trays, count it by hand, and compare.
How does it split a lane stop from a grader stop?Take one real lane jam and see what the grader's OEE shows.
Who gets blamed when the palletizer stops?Pick one real stop and ask which machine the report names first.
Where do downgrades go?Ask how line-made breakage is kept apart from eggs that arrived cracked.
Who can watch the video, and is it kept?Ask for the retention rule and the access list in writing.
What does a fair pilot look like?One grader, one month, and a weekly hand count at the hand-pack lane. Agree it before you sign.

The guide to running a computer vision OEE pilot has a plan you can adapt.

Privacy at the hand-pack lane

The hand-pack lane and the pallet area are where people work closest to a camera. In many European centres, employee representatives will want a say before it goes up, so bring them in first.

  • Point it at the trays and cases. The count needs the packs, not faces.
  • Put the video rule in writing. Video is streamed on request and not stored in Fabrico.
  • Name the purpose. Counting output and giving video of stops, never scoring the people on the lane.

Automatic masking of people in view is roadmap work. For the wider approach, see production line cameras without surveillance.

How Fabrico does it in an egg packing centre

Fabrico is a cloud OEE and light MES platform. Here is what it does in an egg packing centre today, and what it leaves to others.

  • Grader and packers: stops and counts come from the machine signal, such as the grader's run signal and egg count where the controller shares them, or a sensor on a packer. The operator adds the stop reason on a tablet or phone.
  • OEE and losses: by line and shift, with losses shown in money, on the web, iOS and Android.
  • Camera add-on: AI cameras count output, including at a hand-pack lane with no signal. Open the camera video for the minute of a stop from the shift timeline, where a camera covers the station. Video is streamed on request and not stored in Fabrico.
  • Data out: REST API, webhooks, and Excel or CSV export.
  • Roadmap: camera-detected stops, jam zones, and a camera that proposes a likely stop reason for the operator to confirm. None of it is live today.
  • Not Fabrico: crack, dirt, leaker and blood detection, weighing and grading, and checks on printed codes. They stay with the grader and its maker.

For a wider look at what the platform covers, see the Fabrico OEE overview.

Book a demo.

What egg packers ask next

Can Fabrico's camera detect cracked or dirty eggs?

No. Crack, dirt, leaker and blood detection are inspection jobs for the grader's own modules, supplied by the grader maker. Fabrico's camera sits downstream: it counts packs and cases; where it covers a station, the video for the minute of a stop can be opened from the shift timeline, streamed on request and not stored in Fabrico.

Will a camera disturb the grader's own inspection?

It should not, as long as it hangs over the packers, the hand-pack lane or the pallet area, clear of the inspection modules and their lighting. Agree the mounting points with your grader maker first.

Does every packing lane need a camera?

Usually not. Automatic packers have signals, so start with the stations that have none, like a hand-pack lane, and with the pallet area, where stops get blamed on the wrong machine.

Can the camera check the code printed on each egg?

No. Checking printed codes is inspection, so ask your grader maker or printer supplier. The OEE camera counts packs and shows the video for stops.

Does Fabrico keep the camera video?

In the standard Fabrico setup, video is streamed on request and not stored in Fabrico. Ask every vendor the same question, and get the answer in writing for your employee representatives.

What drives the cost of computer vision in an egg packing centre?

The number of graders, lanes and cameras, the mounting, the network in the egg room, and training. Send every vendor the same station list, with your grader model, so the quotes line up.

Book a demo.

Track OEE and downtime in real time
See where lines lose time
Book a live demo.
See where lines lose time
See where lines lose time

Real Fabrico tiles, sample data.

By clicking the Accept button, you are giving your consent to the use of cookies when accessing this website and utilizing our services. To learn more about how cookies are used and managed, please refer to our Privacy Policy and Cookies Declaration