Gridbots MattressSCAN is a perfect combination of technologies which utilises Robotics + AI + Machine Vision together to solve a complex problem: consistent, objective, 100% end-of-line quality inspection of mattresses.
The system can scan multiple sides (Top + Bottom + All 4 Sides) for fabric defects and can also measure the dimensions of the mattress including height. The system generates a 3D profile along with 2D camera based images.
The system utilises a high speed GPU based compute mechanism to analyse the images in real time.
The scanning of each mattress can be completed within 10 seconds.
MATTRESS INSPECTION SYSTEM
Precision Servo Conveyors
0.1 MM Scanning Resolution
Bed Size:[Length: 8 Meters / Width: 3 Meters ]
4 Cameras system
Automatic Accept / Reject Signalling
Barcode / QR Based Traceability
Auto Learn Feature
Programmable Speed of Scanning
Fabric+Width+Height Data at multiple points
Fast ROI and High Speed Operation
Image Archive of Every Mattress Scanned
Shift / Batch Wise Quality Reports
Each mattress passes through a fully automatic, four-stage inspection sequence. No operator judgement is involved in the accept / reject decision, so every mattress is held to the same standard on every shift.
| General | |
|---|---|
| Model Series | GB-MV-MI/2000 |
| Mattress Size Supported | LxBxH [2500x2000x400 mm] |
| Inspection Bed Size | 8 m (L) x 3 m (W) |
| Surfaces Inspected | Top + Bottom + 4 Sides |
| Cycle Time | Upto 10 Seconds |
| Mattress Types | Pocket / Bonnell Spring, PU Foam, Memory Foam, Latex, Coir, Hybrid |
| Vision & Measurement | |
| Camera Count and Types | Laser Profilometer + 2D Stereo Camera (4 Camera System) |
| Scanning Resolution | 0.1 mm |
| Linear Motion Resolution | 1 mm |
| Height Measurement Resolution | 1 mm |
| Data Captured | Fabric image + Length + Width + Height at multiple points + 3D profile |
| Compute | Industrial PC with GPU for real time AI inference |
| Motion & Control | |
| Conveyor | Precision Servo Conveyor, programmable speed |
| Motion Controller | gridMOVE |
| User Defined Recipes | Upto 1000 Supported |
| Operator Interface | Touch screen HMI with live defect overlay |
| Line Interface | Digital I/O for Pass / Fail / Reject; Ethernet for data export |
| Electrical & Environment | |
| Power Rating | 500 Watts |
| Power Consumption | Less Than 500W |
| Supply | 230 V AC, 1 Phase, 50 Hz |
| Operating Temperature | 0 – 45 °C |
| Compressed Air | Not Required |
MattressSCAN combines 2D AI inspection for surface defects with 3D laser profiling for geometric defects. The defect library is extendable through the Auto Learn feature using your own good / bad samples.
| Defect Category | Examples | Detection Method | Surfaces |
|---|---|---|---|
| Fabric Defects | Holes, cuts, tears, snags, pulled threads | 2D AI Vision | All 6 |
| Stains & Contamination | Oil / grease marks, dirt, glue spots | 2D AI Vision | All 6 |
| Stitching Defects | Open seams, skipped stitches, loose threads, wavy quilting | 2D AI Vision | Top, Bottom, Sides |
| Tape Edge Defects | Tape edge misalignment, open / puckered tape | 2D + 3D | 4 Sides |
| Label & Branding | Missing, skewed or wrong label | 2D AI Vision | Configurable |
| Dimensional Defects | Out of tolerance length, width, height | 3D Laser Profile | Full body |
| Shape Defects | Uneven height, crown / dip, bulges, non-square corners | 3D Laser Profile | Top, Sides |
Full 6-side inspection including dimension check.
Manual figures reflect typical visual inspection at line speed.
Manual inspection throughput falls as fatigue builds through the shift.
AI analysis runs in parallel with conveyor motion.
| Parameter | Manual Inspection | Gridbots MattressSCAN |
|---|---|---|
| Time per mattress (6 sides + dimensions) | 60 – 90 s | ≤ 10 s |
| Bottom surface inspection | Requires flipping, often skipped | Automatic, every mattress |
| Height measurement | Tape, 1–3 points | Laser, full 3D profile |
| Consistency across shifts | Depends on operator & fatigue | Same standard 24 x 7 |
| Image record of each mattress | ✗ No | ✓ Yes, linked to serial no. |
| Defect trend analytics | ✗ Manual logbook | ✓ Automatic reports |
| Customer complaint traceability | ✗ Limited | ✓ Full image + dimension history |
| Inspectors required per shift | 2 – 3 | 0 (1 operator supervises) |
| Measurement | Range | Resolution | Method |
|---|---|---|---|
| Length | Upto 2500 mm | 1 mm | Servo encoder + laser edge detection |
| Width | Upto 2000 mm | 1 mm | Laser profile |
| Height (multi-point) | Upto 400 mm | 1 mm | Laser profile |
| Corner squareness | — | 0.5° | 3D point cloud |
| Surface flatness / crown | — | 1 mm | 3D point cloud |
The savings come from three places: fewer inspectors per shift, fewer defective mattresses reaching customers, and faster resolution of complaints using the stored image record.
| Benefit | Typical Impact |
|---|---|
| Inspection labour | Upto 70% reduction |
| Defects escaping to customer | Upto 80% reduction |
| Warranty / return claims | Significant reduction |
| Line throughput | Inspection no longer the bottleneck |
| Typical payback | 12 – 18 months |
Indicative, for a 2-shift plant. Actual payback depends on volume and labour cost.
Live Defect Overlay on HMI
Recipe Management by SKU / Size
Adjustable Tolerance & Severity Thresholds
User Levels: Operator / Supervisor / Admin
PDF / Excel Shift & Batch Reports
Defect Pareto & Trend Dashboards
Data Export to ERP / MES
Remote Support & Model Updates
| Item | Standard | Optional |
|---|---|---|
| Inspection frame with servo conveyors | ✓ | |
| Laser profilometer + 2D camera set with lighting | ✓ | |
| GPU industrial PC with MattressSCAN software | ✓ | |
| Touch HMI and control panel (gridMOVE) | ✓ | |
| Installation, commissioning & operator training | ✓ | |
| Barcode / QR scanner | ✓ | |
| Automatic reject / diverter conveyor | ✓ | |
| ERP / MES integration | ✓ | |
| Annual maintenance contract | ✓ |
| Can it handle different mattress sizes on the same line? | Yes. Each size / SKU is stored as a recipe (upto 1000) and is selected automatically by barcode or manually from the HMI. |
| How is the bottom surface inspected without flipping? | The conveyor design gives the cameras a view of the underside as the mattress passes, so no flipping or handling is needed. |
| How does the system learn our defects? | With the Auto Learn feature, your quality team marks good and defective samples; the AI model is trained on them and deployed on the system. |
| Can it be added to an existing production line? | Yes. It sits at end of line before packing and connects to existing conveyors via digital I/O. |
| What happens to a rejected mattress? | The system flags it on the HMI with marked defect images, signals the line, and optionally diverts it to a reject conveyor. |
| Is support available after installation? | Yes. Gridbots provides on-site and remote support from India, with periodic AI model updates. |