Returnable beverage-crate pocket inspection
Returnable beverage-crate inspection needs coverage of internal pockets and relevant structural features. Separate reflection changes from actual obstructions.
Crate inspection is a camera check of every crate on a returnable bottle line: before the uncaser it finds foreign or wrong bottles and caps left on, after the washer it confirms empty crates are empty and clean, and after the caser it confirms every position holds a capped bottle. addaScan builds it on a HIKROBOT camera above the crate conveyor with a crate diverter. An indicative installed budget for a typical application is R 250 000 – R 1 370 000 excluding VAT, depending on 3D and deep learning.
Returnable glass bottles are a large part of beer and soft drink packaging in South Africa. HEINEKEN South Africa reported investing R2.3bn in its returnable bottling programme here (HEINEKEN, 2024), Coca-Cola Beverages South Africa runs 11 plants across six provinces (CCBA), and its newest returnable is a 2 litre refillable PET bottle, not glass (CCBA, 2025). A returnable bottle system brings crates back from the trade with whatever the trade put in them: bottles of other brands and sizes, bottles with caps still on, straws, cigarette ends and broken glass.
Each of those costs something downstream. A foreign bottle jams the uncaser or the filler, a capped bottle passes the washer without being cleaned, and debris in an empty crate ends up next to clean bottles. The food hygiene regulations require that a container used for food is clean and treat foreign matter and glass as contamination (R638, 2018). The rule sets the outcome, not the method; crate inspection is one way to keep foreign, capped and damaged bottles away from the washer and the filler, crate by crate.
Crate inspection does not replace the empty bottle inspector after the washer, which looks inside each bottle. It works one step earlier and one step later, on whole crates.
| What the line reports | What a station checks | Where it stops |
|---|---|---|
| “Foreign bottles keep jamming the uncaser” | Bottle type in each position of the full crate of empties: neck finish, diameter and colour seen from above, compared with the bottles taught for this crate. | A foreign bottle that looks the same from above, with the same finish and colour, is not separated. |
| “Bottles going into the washer with caps still on” | Crown or screw cap still on a returned bottle, from the top view. | A cap pushed inside the bottle is not seen from above; the empty bottle inspector looks inside. |
| “Short crates going out to the trade” | Full crate inspection after the caser: every position holds a bottle, and every bottle has a cap. | Fill level is not checked here; that is done at the filler. |
| “Straws, cans and rubbish in the crates” | Empty crate inspection after the uncaser: any object left in the bottle crate, and crate type and colour. | Liquid residue on the crate floor and a dirty crate wall are hard to see from above alone. |
| “Wrong-size bottles in the crate” | A 3D height check or a side view over the crate wall finds bottles taller or shorter than the crate type. | Without height data, bottles of similar finish but different height are confirmed on samples only. |
One camera looks down into each crate as it passes on the crate conveyor. The vision software compares each bottle position with what the crate should hold and signals the PLC, which diverts the crate to a manual sorting lane.
Swipe sideways to see the whole drawing →
| Component | HIKROBOT example | Role in this station |
|---|---|---|
| Camera | HIKROBOT CS-series GigE area scan camera, 5–20 MP mono or colour (e.g. MV-CS050-10GC, MV-CS200-10GM), or a CT-series camera rated IP67 with lens cap where washer spray reaches it | Images every bottle position in the crate from above |
| Processing | HIKROBOT VC3000 vision controller running HIKROBOT VisionMaster; for a fixed crate layout an SC5000X or SC6000 smart camera can replace the controller | Runs the crate recipe and sends the result to the PLC |
| Lighting | Two HIKROBOT bar lights or a large flat light above the crate, with a shroud | Even light into every position; hides the plant lights |
| Trigger | HIKROBOT MV-PE photoelectric sensor or the existing crate sensor | Fires the camera when the crate is in position |
| Height check (option) | HIKROBOT DB-series 3D stereo camera (e.g. DB500S) or a DP-series 3D laser profiler | Measures bottle height across the crate for wrong-size bottles |
| Reject | Not HIKROBOT: pneumatic pusher or diverter, from the conveyor supplier | Moves failed crates to a sorting lane |
| Confirmation and enclosure | Local supply: confirmation sensor, stainless camera bracket, splash cover | Proves the crate left; protects the camera near the washer |
| Tool or software | What it does here |
|---|---|
| HIKROBOT VisionMaster pattern and blob tools | Find each bottle position and decide whether a bottle is present. |
| VisionMaster deep-learning classification | Sort bottles by type where brands differ by finish, colour or embossing; trained on your bottles. |
| VisionMaster anomaly detection | Flag anything unusual in an empty crate, trained on good empty crates only. |
| Recipe per crate type | One recipe per crate and bottle combination, selected by the PLC at changeover. |
| Result and image log | Reject reason per crate and saved images of rejects for the line team. |
Camera and light on a stainless bridge over the crate conveyor, high enough for full crates to pass; IP67 camera or splash cover where washer spray or condensation reaches.
A light as large as the crate and a shroud around the camera bridge, so wet glass reflects the same light on every shift.
One trigger per crate; the PLC tracks each crate from camera to diverter so the right crate is moved.
Result, reject reason and recipe number over PROFINET or EtherNet/IP where the chosen HIKROBOT device lists it, or digital I/O on simple stations.
Diverter with a confirmation sensor and a full-lane signal; if the vision system is not ready, crates are diverted or the line is held, not passed unchecked.
Crate and bottle recipes chosen by the PLC at brand changeover; a screen shows the last rejects with the reason.
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.
The low figure is a 2D station that counts missing bottles and flags objects with rule-based tools and signals an existing crate rejector. The high figure adds a 3D camera for bottle height, deep-learning classification of foreign bottles and a new crate pusher with confirmation; 3D work sits outside the standard ranges on the cost guide, which is why it is priced here separately. Changes to your conveyor, utilities and VAT are excluded.
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 83 800 | R 513 600 | Low: one 20 MP mono camera, lens, two bar lights, VC3000 controller. High: 20 MP colour camera, longer lights, 3D camera for bottle height, larger controller |
| Software licence | R 33 800 | R 150 000 | HIKROBOT VisionMaster standard licence; high case with the deep-learning tier |
| Local materials and reject | R 40 000 | R 180 000 | Low: frame, shroud, cabling, trigger, signal to the existing crate rejector. High: stainless frame and light tunnel, crate pusher with confirmation and lockable lane, panel, UPS, screen |
| Engineering (recipes, PLC link) | R 50 400 | R 260 000 | 56 h at R 900 to 200 h at R 1 300 |
| Installation, FAT, SAT, training, documents | R 28 800 | R 124 800 | 32 h to 96 h |
| Project management | R 14 400 | R 62 400 | 16 h to 48 h |
| Travel (Gauteng) | R 3 000 | R 10 000 | Outside Gauteng quoted separately |
| Total (rounded) | R 250 000 | R 1 370 000 | Low: one view. High: with 3D height and deep learning |
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 |
|---|---|---|---|
| Area scan cameras with a vision controller | Core | One overhead area scan camera covers a whole crate; a controller handles several crate recipes. | Camera and lens selection |
| Presence detection and counting | Core | Every bottle position is a presence check: missing bottles in full crates, leftover items in empty ones. | Presence and counting |
| Deep learning (classification, segmentation, DL OCR) | Supporting | Classification separates foreign bottles that differ in colour or embossing; anomaly detection flags unexpected objects. | Rule-based vs deep learning |
| 3D laser profiling | Supporting | Wrong-height bottles need height data that a single top view cannot give. | Dimensional inspection |
| Lighting technique | Supporting | Wet glass and crate walls create reflections and shadows; a large, shrouded light keeps the image stable. | Machine vision lighting |
| PLC link and reject with confirmation | Supporting | Crates are heavy: diverting the right crate needs tracking and a confirmed diverter. | PLC and reject integration |
Every station is accepted on your products, against numbers agreed in writing before the build. How we deliver →
Crates of your bottles plus the foreign bottles, caps and rubbish you see from the trade, imaged with the proposed camera and light; the result says what is reliably separated and what is not.
Before delivery, a set of marked challenge crates (missing bottle, foreign bottle, capped bottle, item in an empty crate) must each be diverted, and good crates must pass.
On your line, at production speed, the challenge crates are run again, and false diverts and escapes are counted over an agreed run of production.
A top view of a crate answers which visible positions are occupied and whether the exposed bottle features match the expected crate. It does not inspect every bottle wall or base.
| Decision or event | Check or action | Acceptance boundary |
|---|---|---|
| Position map | Check each expected crate cell | Occluded cells cannot be counted by assumption. |
| Wrong bottle type | Compare visible discriminating features | Different bottles may share a similar top view. |
| Remaining cap or debris | Check the visible reject condition | Hidden contamination remains outside the view. |
| After washing | Use a separate individual-bottle check where required | A complete crate is not proof that washed bottles are acceptable. |
Associate the crate result with its own trigger and route uncertain crates to controlled inspection. Keep the before-washing crate task separate from the after-washing bottle task.
Crate inspection is a camera check of whole crates of returnable bottles. Before the uncaser it finds foreign bottles, wrong bottles and bottles with caps still on; after the uncaser it confirms empty crates are empty; after the caser it confirms every position holds a capped bottle. Failed crates are diverted to a sorting lane instead of stopping the line.
Often, yes, when the bottles differ in neck finish, diameter, colour or embossing seen from above. Deep-learning classification trained on your bottles and the foreign bottles you actually receive handles the harder differences. Bottles that look identical from above, with the same finish, colour and height, cannot be separated by a top view; a 3D height check helps only when the height differs.
An indicative installed budget is R 250 000 – R 1 370 000 excluding VAT, at 2 October 2026 prices. The low figure is a one-camera 2D station signalling your existing crate rejector; the high figure adds a 3D height check, deep-learning bottle classification and a new crate pusher with confirmation. Both include hardware, software, engineering, installation, FAT and SAT in Gauteng. A firm price follows a sample test on your crates. This is a typical application: every project involves custom requirements, so the final price depends on your products, line and acceptance criteria.
No. Crate inspection works on whole crates before the washer and after the caser. An empty bottle inspector looks into each bottle after the washer for residue, chips and foreign objects inside. Lines that reuse glass usually need both; crate inspection reduces what the washer and the empty bottle inspector have to deal with.
The same station principle applies: bottle presence, caps and foreign bottles are checked from above. PET reflects light differently from glass and the crate layout differs, so PET crates get their own recipe and are included in the sample test.
Usually in two places: on the full crates of empties before the uncaser, to divert crates with foreign or capped bottles, and on the full crates after the caser, to stop short crates going to trade. An empty crate check after the uncaser, before the crate washer, is a third option.
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.