China Automated Vision Inspection Solutions

Glass Print Inspection Machine
Applicable Defect Inspection:
- LOGO Print Inspection
- Print Defect Inspection
- Glass Blank Margin Dimension Inspection
- Architectural Glass
- Automotive Glass
- Display Glass
- Home Appliance Glass
Overview
The glass online inspection system can detect defects in real-time during production and provide product quality information. It records quality data to optimize production processes. The system performs non-contact edge defect detection, surface flaw inspection, and dimensional measurement for products of varying widths and machine speeds.
Using a linear image sensor, the vision system scans targets in real-time to form images. Specialized image processing and analysis software based on an industrial control computer automatically detects surface defects, measures dimensions, and logs defect data into a categorized database. The system features a user-friendly Windows-based graphical interface.
Highlights
-- Recipe Management
The software allows customized parameter settings for each product type, saved as a “recipe.” During production, simply selecting the recipe automatically loads the corresponding parameters.
-- Classification
Defects are accurately classified using AI-based analysis of captured images, enabling customized criteria for different defect types.
-- Defect Recording (Database Functionality)
Defect characteristics and image data are stored in database files for offline queries, reporting, and printing.
-- Real-Time Interface Display
Instantly shows detected defect images, characteristic data, and trend charts.
Defects Classification
-- LOGO Print Defects
Dot missing, Ink leakage, Pattern missing, Pattern omission, Pattern blur, A-hole blockage, Position deviation.
-- Print Defects
Pin holes in printed area, Feature offset, Ink leakage in visible area, Contamination.
-- Glass Blank Margin Defects
Transition point missing, Transition point bridging, Edge jaggedness, Ink scratches.
Machine Performance Specification
Parameter | Value |
---|---|
Glass Type | Flat Glass |
Physical Resolution | 0.042 mm |
Power Supply Voltage & Power Rating | 220V, 2kW |
Ambient Temperature | Below 30°C |
Detection Target Range (Conveyor Dir.) | Field adaptable width |
Min. Roller Gap at Detection Station | 50 mm |
Max. Line Speed | 40 m/min |
Fastest Detection Rate (excl. scan) | < 3 sec/piece |
Detection Capabilities
Detection Location | Detection Item | Detection Capability | Remarks |
---|---|---|---|
Positioning Holes | Missing hole | 100% detection rate | |
Light leak (brightness settable) | >0.15mm | ||
Contamination | Not detectable | Invisible under backlight | |
Positioning hole offset | 100% detection rate ≥1mm | ||
Printed Area | VIN frame offset | 100% detection rate >1mm | |
Camera frame jaggedness | Not detectable | ||
MB frame jaggedness | Not detectable | ||
Color difference | Not detectable | Invisible under backlight | |
Water marks | Not detectable | Invisible under backlight | |
Ink leakage | 100% detection rate | Ink leakage or contamination. Capability >0.15mm (software configurable) | |
Non-Printed Area | Contamination | >0.15mm | OK judgment based on set quantity threshold |
Blank Margin | Upper edge margin excess | 100% detection rate >0.15mm | Measured from 4 edges. Product requirement ≥0.5mm, equipment capability 0.15mm |
Lower edge margin excess | 100% detection rate ≥0.15mm | ||
Left edge margin excess | 100% detection rate ≥0.15mm | ||
Right edge margin excess | 100% detection rate ≥0.15mm | ||
Logo | Character/pattern omission | 100% detection rate | Detectable even for missing dot-shaped characters |
Character/pattern missing | Missing single char/pattern >20% | Counts independent characters/pattern areas | |
Character/pattern blur | Detectable | Visually apparent | |
Antenna / Heater | Wire breakage | Detectable | Visually apparent; capable of detecting breaks in small heater wire areas. Data to be provided after algorithm optimization & testing. |
Transition Points | Missing/bridging points | Not detectable | Current algorithm can compute but requires long processing time (≥10min). Optimization planned for future projects. |
Defects Detected During Projects
-- LOGO Defects

-- Ink Overflow

-- Pin Holes

-- Blank Margin Defects
