Real-Time Molten Pool Defect Detection Technology for Overlay Welding Using Industrial Cameras

In modern industrial manufacturing, overlay welding (cladding) is widely used across machinery, energy, chemical processing, aerospace, and other industries for component repair and surface performance enhancement. However, traditional quality control in overlay welding relies heavily on operator experience or post-weld inspection, making real-time monitoring difficult. As a result, defects are often detected too late, leading to reduced productivity and increased costs.

Today, let’s explore MindVision industrial camera–based real-time molten pool defect detection technology for overlay welding processes, and see how intelligent visual monitoring enables higher welding quality and significantly improved production efficiency.

Technology Principles

High-Speed Imaging & Dynamic Exposure Control

Molten pool monitoring cameras typically utilize high-speed CMOS sensors with frame rates reaching 200–500 frames per second, enabling accurate capture of rapid molten pool dynamics.
Combined with dynamic exposure control and high optical density filters, the system effectively suppresses arc glare and spatter interference, ensuring clear and usable images even under extreme high-contrast welding conditions.

Key Technical Advantages

1. High-Precision Molten Pool Dynamic Imaging

By integrating high-speed industrial cameras with High Dynamic Range (HDR) imaging, the system clearly captures molten pool geometry, thermal distribution, and flow behavior.
It operates reliably in harsh welding environments with strong arc light and spatter, delivering stable and interference-free images.

2. Real-Time Intelligent Defect Detection

Powered by AI-based deep learning algorithms, the system analyzes molten pool images in real time to automatically identify common welding defects such as porosity, slag inclusion, lack of fusion, and undercut.
Defect evaluation criteria can be customized to meet different welding processes and application requirements.

3. Instant Alarm & Process Optimization

When defects are detected, the system immediately triggers visual/audio alarms or machine stop signals, preventing large-scale production of defective parts.
By combining historical data analysis, it also provides process optimization recommendations, helping improve welding stability and consistency.

4. Complete Data Traceability

All welding process data—including molten pool images, defect types, and timestamps—are fully recorded, enabling comprehensive quality traceability and post-process analysis.

Application Cases

Heavy Equipment Overlay Welding

On a heavy construction machinery cladding production line, molten pool cameras monitor layer thickness in real time. With closed-loop control, rework rates were significantly reduced, resulting in a substantial increase in productivity.

Aerospace Critical Components

For additive repair of aerospace turbine blades, the system achieved micron-level molten depth measurement, enabling precise control of cladding geometry. Porosity and cracking defects were dramatically reduced, meeting stringent aerospace-grade quality standards.

Extended Industrial Applications

The technology also shows strong potential in petrochemical reactor vessels, pipeline cladding repair, rail transit wheelsets, and track welding maintenance.

Conclusion

Real-time molten pool defect detection technology for overlay welding—powered by high-speed industrial cameras and advanced image processing—enables comprehensive monitoring and closed-loop quality control throughout the welding process. It brings unprecedented improvements in efficiency, consistency, and reliability to the welding industry.

As the technology continues to mature and costs decrease, it will see broader adoption across heavy industry, energy, and aerospace sectors, accelerating the transition toward intelligent and precision-driven welding manufacturing.

You may contact us at chenguo@mindvision.com.cn to gain more in-depth technical insights and practical applications in the fields of machine vision and optical imaging.