Rolling Defects: How to Detect Them Before Finishing

In the metal processing industry, the rolling process is a critical step to achieve the desired dimensional and mechanical properties of materials. However, it is also during this phase that defects can originate—defects that, if not detected early, can compromise product quality and lead to significant costs.

Identifying defects before the finishing stage is therefore essential to avoid rework, scrap, and downstream production issues. Today, thanks to Computer Vision and Artificial Intelligence, it is possible to detect these anomalies automatically and in real time.

What Are Rolling Defects?

During rolling, metal is plastically deformed through rollers to achieve the required thickness and shape. Several types of defects can occur in this process.

The most common include:

• surface cracks

• inclusions

• folds and overlaps

• roll marks

• flatness defects

• oxidation or contamination

These defects may be difficult to detect in early stages but become more visible and problematic during finishing—or even worse, in the final product.

Why Detect Defects Before Finishing

The finishing phase (polishing, coating, surface treatment) is where the product gains significant added value.

If a defect is detected only at this stage:

• scrap costs are much higher

• time and resources are wasted

• production efficiency decreases

• the risk of delivering non-compliant products increases

Detecting defects upstream, immediately after or during rolling, allows manufacturers to take corrective action quickly and minimize losses.

Limitations of Traditional Inspection Methods

Historically, rolled products have been inspected through:

• manual visual inspection

• traditional optical systems

• sample-based checks

These approaches have several limitations:

• difficulty handling large production volumes

• dependency on operator experience

• inability to detect micro-defects

• delayed identification of anomalies

On high-speed production lines, these methods are no longer sufficient.

How Computer Vision Detects Rolling Defects

Computer Vision systems enable real-time surface inspection directly during the rolling process.

The system typically operates through several stages.

1️⃣ Image Acquisition

High-speed industrial cameras are installed along the rolling line.

With advanced lighting systems (laser, structured light), it is possible to highlight:

• surface imperfections

• texture variations

• geometric defects

2️⃣ Image Analysis

Vision algorithms continuously analyze images to detect anomalies.

The system can identify:

• linear defects (cracks, marks)

• point defects (inclusions, contamination)

• surface variations

• flatness issues

With AI and Deep Learning, the system can also detect complex or previously unseen defects.

3️⃣ Defect Classification

Once a defect is detected, the system classifies it based on:

• type

• severity

• location

This enables fast and informed decision-making during production.

4️⃣ Immediate Action

The system can automatically trigger actions such as:

• rejecting defective material

• alerting operators

• adjusting rolling parameters

• recording data for further analysis

Benefits of Early Detection

Implementing Computer Vision in rolling processes provides significant advantages.

• Reduced Scrap

  Defects are detected before finishing, avoiding high-value losses.

• Increased Production Efficiency

  Rework and process interruptions are minimized.

• 100% Inspection Coverage

  Every meter of material is analyzed in real time.

• Process Improvement

  Collected data helps identify root causes of defects quickly.

• Higher Product Quality

  Only compliant material continues through the production process.

Industrial Applications

Computer Vision for rolling defect detection is widely used in:

• steel rolling

• aluminum production

• strip and coil processing

• automotive industry

• structural component manufacturing

In all these applications, material quality is critical.

The Future: Intelligent and Predictive Inspection

Advances in AI and analytics are driving the development of increasingly advanced inspection systems.

In the future, manufacturers will be able to:

• predict defect formation

• correlate anomalies with process parameters

• integrate inspection with MES systems

• automatically optimize production

This will lead to smarter and more autonomous rolling lines.

Conclusion

Rolling defects represent one of the main challenges in metal processing, especially when detected too late.

Thanks to Computer Vision, manufacturers can now detect these defects in real time before the finishing stage, reducing costs, improving efficiency, and ensuring higher product quality.

In an increasingly competitive industrial environment, early defect detection is not just a technical improvement—it is a strategic advantage.

Want to know more? Contact us at info@metalya.it