Stitched maize-meal bag-mouth inspection
A continuous stitch line does not prove that a maize-meal bag mouth is closed correctly. Reject on relative geometry.
Line solution
Test this on your samplesMaize meal bag inspection: print, code and closure
// Short answer
Maize meal bag inspection is a camera check of every filled bag after the closer: the printed product name, grade and mass belong to the product being packed, the date and batch code are present and readable, and, with a second camera, the bag is properly closed. addaScan builds it on HIKROBOT cameras in dust housings with a bag reject, and the same station suits sugar and flour. An indicative installed budget for a typical application is R 110 000 – R 770 000 excluding VAT.
// At a glance
The station in three lines
- Faults caught
- Wrong bag stock for the grade or size; missing or unreadable date and batch code; bag not closed
- Configuration
- HIKROBOT vision sensor or two area cameras in dust housings, bag counting, bag reject with confirmation
- Line link
- Digital I/O, or PROFINET or EtherNet/IP where the chosen HIKROBOT device lists it, to the filler, closer or palletiser PLC
// Fit & limits
What this station will not do, and what it needs
What it will not do
- Mass: the camera reads the mass printed on the bag; it does not weigh the bag. That is the checkweigher.
- Closure strength: it sees that the bag is sewn or sealed across the top, not whether the closure holds or leaks meal.
- Inside the bag: contents, foreign matter and infestation are not seen.
- Code grading: reading a code or barcode on the bag is not grading its print quality.
What makes it work
- Bags presented the same way: face up and in the same orientation at the camera.
- Samples of every bag stock: good bags of each grade and size, and bags with each fault you want caught.
- Dust routine: air purge running and housings wiped on a schedule.
- Room to reject: a pusher or flap that can move a full bag, and a lane for it.
// On South African lines
Maize meal bags in South Africa: what the marking rules ask for
Maize meal is the main staple food in South Africa: the National Agricultural Marketing Council calls bread the second most important staple after maize meal (NAMC). Mills pack it in pre-printed maize meal packaging bags of several grades and sizes, often on the same filling line, with a change of bag stock at each grade or size change and a date and batch code printed on each bag as it is filled.
The maize products regulations require each container to be marked in clearly legible symbols, letters and figures with, among other things, the name and address of the packer, the grade (such as super or special) and the net mass, and they require bags to be “filled in accordance with the mass depicted and properly closed” (Maize products regulations, 2016). R146 adds a batch number on every container, so the batch is identifiable and traceable (R146, reg 11). A bag of special filled into super bag stock, a code that the printer missed, or a bag that leaves the closer open, each fails one of these.
The rules set the outcome, not the method. A camera is one way to check the marking and the closure on every bag. The mass in the bag is a different check: the Legal Metrology Regulations require packers using automatic filling or weighing instruments to take steps to detect incorrect quantities (Legal Metrology Regulations, 2018), which is the checkweigher’s job.
What the mill reports, and what the station checks
| What the line reports | What a station checks | Where it stops |
|---|---|---|
| “Special in super bags after the changeover” | The pre-printed bag printing for this run, product name, grade and mass, is matched against the recipe by pattern and text on every bag. | If the recipe itself is set to the wrong grade, the station agrees with it; the recipe change is part of the changeover check. |
| “The inkjet stopped and nobody noticed” | Date and batch code present on every bag, read by OCR and compared with the code expected for the shift. | A code printed on the back of the bag, or under a fold, needs its own camera. |
| “You can’t read the code on the creased bags” | Broken, smeared or partly missing characters, read with deep-learning OCR and defect detection trained on your codes and bags. | Codes across a deep crease or on a badly folded bag may stay unreadable; these are rejected, not guessed. |
| “Bags coming off the closer open” | With the second camera after the bag sewing machine or heat sealer, the closure is present across the full bag top. | The camera sees the stitch line or seal from outside; it does not test whether the closure holds or is sift-proof. |
| “The pallet count doesn’t match” | Every bag that passes is counted from the trigger, and rejected bags are counted separately. | Bags added or removed by hand after the station are not counted. |
// The complete solution
Maize meal bag inspection station: hardware, software and installation
Filled, closed bags leave the closer on the take-away conveyor. One camera looks at the printed face; in the fuller station a second looks at the closed top. Both sit in dust housings with air purge. A failed bag is pushed or flapped off before the palletiser.
[Schematic]Plan view, not to scale
Swipe sideways to see the whole drawing →
[1]: A photoelectric sensor fires the cameras and counts bags.
[2]: Looks at the printed face: name, grade, mass, date and batch code.
[3]: In the fuller station, sees the sewn or sealed top.
[4]: A pusher or flap removes failed bags.
[5]: Confirms the failed bag left the line.
Hardware
| Component | HIKROBOT example | Role in this station |
|---|---|---|
| Face camera | Low: HIKROBOT SC3000 vision sensor MV-SC3050M with OCR and character comparison. High: HIKROBOT MV-CS200-10GM 20 MP mono area scan camera with an MVL-HF2524M-10MP lens | Reads the printed name, grade, mass and the date and batch code |
| Closure camera (high) | A second MV-CS200-10GM camera with an MVL-HF2524M-10MP lens | Checks the sewn or sealed top of each bag |
| Processing (high) | HIKROBOT VC3000 vision controller (MV-VC3302P) running HIKROBOT VisionMaster with the deep-learning licence | Runs deep-learning OCR and closure checks for both views |
| Lighting | HIKROBOT MV-LLDS-H-600-30-W bar light (low); two MV-LLDS-H-1400-40-W bar lights (high) | Even light across the full bag width and the top seam |
| Dust housing | Not HIKROBOT: IP67 camera housing per camera, with an air purge | Keeps flour dust off the lens; HIKROBOT lists no dust or air-purge housing |
| Trigger | HIKROBOT MV-PE photoelectric sensor | Fires the cameras once per bag and counts bags |
| Reject | Not HIKROBOT: bag pusher or flap with confirmation sensor (high); low case signals the existing reject or stops the conveyor | Removes failed bags before the palletiser |
Software and vision tools
| Tool or software | What it does here |
|---|---|
| HIKROBOT SCMVS OCR and character comparison (low) | Read the date and batch code and the printed grade and mass, and compare them with the recipe. |
| VisionMaster deep-learning OCR (high) | Read characters on creased paper or woven plastic, with low contrast and deformation. |
| VisionMaster deep-learning defect detection of printed codes | Flag inkjet codes with broken, smeared or missing characters, trained on your bags. |
| VisionMaster pattern and edge tools (high) | Match the pre-printed bag panel to the run and find the stitch line or seal across the bag top. |
| Recipe per product and bag | Grade, mass, code layout and closure type per product, selected at changeover. |
Installation and integration
01
Mounting and enclosure
Cameras in sealed housings on rigid mounts over the take-away conveyor, away from the filling spout; mounts strong enough for a mill environment and not shaken by bag drops.
02
Lighting and shrouding
Long bar lights across the bag width, angled so glossy plastic bags do not glare; a shroud keeps the light the same on every shift.
03
Trigger and tracking
One trigger per bag, which also counts; each bag is tracked from the camera to the reject so the right bag is removed.
04
PLC link
Result, reject reason and bag count by digital I/O, or over PROFINET or EtherNet/IP where the line PLC uses it and the chosen HIKROBOT device lists it.
05
Reject and confirmation
Pusher or flap with a confirmation sensor and a full-lane signal; if the vision system is not ready, bags are rejected or the line is held, not passed unchecked.
06
Recipes and operator screen
Recipe chosen at grade or bag-size changeover; a screen shows the last rejected bags with the reason, and a dust-cleaning reminder for the housings.
// Indicative cost in South Africa
What this station costs, excluding VAT
Indicative budget for a typical application of this kind. Every project involves custom requirements (your products, line speed, layout, PLC and acceptance criteria), so the final price differs from case to case. A firm price follows the sample test.
- Low
- R 110 000
- one view of the bag face
- High
- R 770 000
- face and closure, deep learning, own reject
- Basis
- Typical application of this kind, indicative
- South African rand, excluding VAT
- Prices as at 2 October 2026
- Travel (Gauteng): outside Gauteng quoted separately
// Included and excluded
The low figure is one vision sensor in a dust housing that reads the date and batch code and checks the printed grade and mass, counts bags from the trigger and signals an existing reject. The high figure uses two cameras on a controller, one on the bag face and one on the closure, with deep-learning OCR, air-purged housings and a bag reject with confirmation. Changes to your conveyor, compressed air supply, utilities and VAT are excluded; a checkweigher is not part of this station.
Hardware is priced from public list prices converted at the ECB reference rate, plus freight, clearance and margin; engineering at R 900 – R 1 300 per hour. How these ranges are built →
Swipe the table sideways to see the high figure and the basis →
| Cost block | Low | High | Basis |
|---|---|---|---|
| Vision hardware (landed, with margin) | R 55 400 | R 183 700 | Low: one SC3000 vision sensor, bar light, IP67 housing. High: two 20 MP mono cameras with lenses, two long bar lights, VC3000 controller, two IP67 housings |
| Software licence | R 0 | R 120 000 | Low: none, the vision sensor software is free. High: HIKROBOT VisionMaster licence with deep learning |
| Local materials and reject | R 15 000 | R 120 000 | Low: bracket, air purge, cabling. High: dust-proof mounts and air purge, bag reject flap or pusher with confirmation, panel |
| Engineering (recipes, PLC link) | R 27 000 | R 182 000 | 30 h at R 900 to 140 h at R 1 300 |
| Installation, FAT, SAT, training, documents | R 14 400 | R 83 200 | 16 h to 64 h |
| Project management | R 5 400 | R 36 400 | 6 h to 28 h |
| Travel (Gauteng) | R 2 000 | R 6 000 | Outside Gauteng quoted separately |
| Total (rounded) | R 110 000 | R 770 000 | Low: one view of the bag face. High: face and closure, deep learning, own reject |
// Related technology
Which technologies this scenario uses, and why
The technologies behind this station, how much each matters here, and where each is explained in depth.
| Technology | Role | Why it matters in this scenario | Read more |
|---|---|---|---|
| OCR/OCV of printed codes | Core | The date and batch code, and the grade and mass on the pre-printed bag, are read as text on every bag. | OCR/OCV inspection |
| Vision sensor (HIKROBOT SC1000–SC3000) | Core | One SC3000 vision sensor in a dust housing covers the bag face on simpler lines, with no PC. | Choosing a smart camera |
| Deep learning (classification, segmentation, DL OCR) | Supporting | Creased paper and woven plastic deform characters; HIKROBOT describes its deep-learning OCR as built for deformed and low-contrast characters. | Rule-based vs deep learning |
| Presence detection and counting | Supporting | Every bag is counted at the trigger, which ties the inspection to the pallet count. | Presence and counting |
| Lighting technique | Supporting | Bags are wide and uneven; long bar lights give the same light across the bag and on the seam. | Machine vision lighting |
| PLC link and reject with confirmation | Supporting | Full bags are heavy; the reject needs tracking, a confirmation sensor and a lane. | PLC and reject integration |
// Acceptance
How the station is proven before handover
Every station is accepted on your products, against numbers agreed in writing before the build. How we deliver →
01
Sample test
Filled bags of each grade and size, with good print and with each fault (wrong bag stock, missing code, smeared code, open closure), imaged with the proposed cameras, lights and housings.
02
FAT with challenge bags
Before delivery, marked challenge bags must each be rejected for the right reason, and good bags of each recipe must pass.
03
SAT on the line
On your line at production speed and in normal dust, the challenge bags are run again, and false rejects and escapes are counted over an agreed run of production.
Match the bag grade and batch to the packing order
For staple foods in printed bags, the bag artwork can identify the grade while the variable print identifies the batch. Both need a trusted expected value.
| Decision or event | Check or action | Acceptance boundary |
|---|---|---|
| Bag grade | Check the approved identifying artwork or code | Similar colours alone may not distinguish grades. |
| Batch/date | Compare printed data with the active order | A readable previous batch is still wrong. |
| Bag closure | Inspect the specified visible closure condition | Visible closure does not establish every barrier property. |
| Change of bag stock | Challenge the new stock against the same order | A material reload can introduce a mismatch without a recipe change. |
Record the expected bag identity and observed result separately. Keep net-content checks with the appropriate filling or weighing control.
// FAQ
Questions about maize meal bag inspection
What does maize meal bag inspection check?
It checks each filled bag after the closer: the product name, grade and mass printed on the bag belong to the product being packed, the date and batch code are present and readable, and, with the second camera, the bag is closed across the top. Bags are counted as they pass, and failed bags are removed before the palletiser.
Can the camera check the net mass of the bag?
No. It reads the mass printed on the bag and checks it is the right figure for the product, but it does not weigh anything. Whether the bag holds that mass is checked by a checkweigher or by the filling scale and its records, as the Legal Metrology Regulations expect. The two stations work side by side.
How much does maize meal bag inspection cost in South Africa?
An indicative installed budget is R 110 000 – R 770 000 excluding VAT, at 2 October 2026 prices. The low figure is one vision sensor in a dust housing on the bag face, reading the code and checking the printed grade and mass. The high figure adds a closure camera, deep-learning OCR on a controller, air-purged housings and a bag reject with confirmation. A firm price follows a sample test on your bags. This is a typical application: every project involves custom requirements, so the final price depends on your products, line and acceptance criteria.
Does the same station work for sugar packaging and flour?
Yes, the principle is the same: printed product details, a date and batch code, and a closed bag. Each product has its own marking rules and bag stock, so each gets its own recipe listing what must be on the bag. Fine sugar and flour dust are handled the same way as maize dust, with sealed housings and air purge.
Is this a bag inspection machine we can buy off the shelf?
No. It is a station built on your existing conveyor from standard HIKROBOT cameras, lights and a controller, set up for your bags and your codes, with a reject that suits your line. Checking empty bags for print faults at the bag maker is a different job, done before the bags reach the mill.
Will the cameras cope with the dust in a maize mill?
They sit in IP67 housings with an air purge that keeps a clean window in front of the lens, because HIKROBOT does not list a dust or purge housing of its own. Housings are wiped on a schedule, and the station tells the operator when images get dull. The sample test and SAT are run in normal production dust.
Sources
- NAMC: Our daily bread (bread the second staple after maize meal) (government agency article, checked 2026-10-02)
- Regulations relating to the grading, packing and marking of maize products (Proclamation R3 of 2016), FAOLEX copy (gazette scan, checked 2026-10-02)
- Regulations relating to the labelling and advertising of foodstuffs (R146, 2010), reg 11 (regulation copy, checked 2026-10-02)
- Legal Metrology Regulations (GN 877, 2018), regs 137–138 (gazette, checked 2026-10-02)
- HIKROBOT VisionMaster (deep-learning OCR for low-contrast and deformed characters) (vendor page, checked 2026-10-02)
- HIKROBOT Smart Product Catalog 2025Q4: deep-learning detection of spray code character defects on food packaging, p.65 (vendor catalogue, checked 2026-10-02)
Related:
More line solutions in food and beverage:
Returnable crates →Multipacks →Can ends and dents →Juice box straws →Dairy cups →Bread bag closures →Meat tray labels →Wine labels →Sachets and pouches →Egg cartons and trays →Bottled water (PET) →Food tray seal inspection →Food packaging version verification →Bakery product vision inspection →Returnable glass bottle inspection →
// Next step
Send us a few filled bags, including the ones that went wrong
Describe the line, the products and the fault. We test on your good and faulty samples and reply with a feasibility note and a budget.
Planning this inspection outside South Africa
Use the country guidance to prepare the line data, interfaces and delivery scope for a project assessed from Johannesburg.
Vision Inspection Systems in Zambia
Bag-mouth inspection and counting answer different questions. Project scope, references and Johannesburg delivery for Zambia.
Vision Inspection & Barcode Readers in Uganda
Grain and feed bags: codes and counts that survive line stops.