Defect detection in slab surface: a novel dual Charge-coupled Device imaging-based fuzzy connectedness strategy

Area of Science:

• Materials Science
• Computer Vision
• Manufacturing Engineering

Background:

• Automated surface defect inspection is crucial for quality control in manufacturing.
• Existing machine vision systems have limitations in accurately detecting and delineating surface defects.

Purpose of the Study:

• To develop a robust and accurate surface defect inspection system for continuous casting slabs.
• To overcome the limitations of individual imaging systems by combining their strengths.
• To enable automated and precise defect extraction and delineation in production lines.

Main Methods:

• A dual scanning imaging system combining line array CCD (LS-imaging) and area array CCD laser 3D scanning imaging (AL-imaging).
• Image registration using maximum mutual information to align data from both sensors.
• Image fusion of 2D and 3D information.
• Region of Interest (ROI) localization and delineation using iterative relative fuzzy connectedness (IRFC) algorithm.

Main Results:

• The proposed system effectively integrates complementary information from 2D and 3D imaging.
• Accurate alignment and fusion of images from different sensors were achieved.
• The IRFC algorithm enabled precise delineation of surface defects.
• Experimental results demonstrate competitive performance compared to state-of-the-art methods.

Conclusions:

• The joint imaging scanning strategy offers a feasible approach for advanced machine vision inspection.
• This method provides a powerful tool for ROI delineation in surface defect analysis.
• The system demonstrates potential for improving automated quality control in industrial production lines.