Vision System for Tire Inspection
Information
- Sector: automotive
- Technology:
In the automotive industry, precision and reliability in quality control processes are essential. Vision systems play an increasingly important role across the entire inspection workflow – from raw material inspection and code reading to defect detection.
By combining carefully selected hardware components with advanced image processing algorithms, including deep learning (DL)-based solutions, the system can reliably detect even the most challenging anomalies, such as inconsistencies in DOT markings on dark, light-absorbing tire sidewalls.
System objective
To verify tire technical data by reading information embossed on the tire sidewall and comparing it with production system records.
A key engineering challenge was ensuring high-quality imaging of the tire sidewall, which directly determines the accuracy of data acquisition. The system performs both optical character recognition (OCR) of tire markings and verification of their compliance with data stored in the manufacturing execution system (MES). Equally important is inspection speed – a complete inspection, including data acquisition from both sides of the tire, is completed in approximately 3 seconds, requiring simultaneous image buffering and real-time data processing.
Another challenge was ensuring consistent performance across a highly diverse product range. Tires vary significantly in size and surface characteristics, directly affecting imaging conditions and data quality. The system has been designed to accommodate tires with outer diameters ranging from 400 to 1,000 mm and heights from 90 to 400 mm.
Data read by the system
- Tire manufacturer
- Tire model
- DOT code for production date identification
- EU type approval markings
- Tire width, aspect ratio, radial construction designation, rim diameter, load index, speed rating, and reinforced tire marking
- Country of manufacture
- Run Flat marking
- M+S (Mud + Snow) marking
The average image analysis time is approximately 850 ms, enabling reliable operation in demanding industrial environments. The system supports conveyor speeds of up to 1 m/s while maintaining accurate detection performance. It can read and verify 23 different types of alphanumeric information, demonstrating both high accuracy and advanced technological capabilities.
