Robust Industrial Localization of Dot-Pattern Data Matrix Codes on Metallic Surfaces

Dot-pattern Data Matrix (DM) codes engraved or printed on metallic surfaces are widely deployed in industrial inspection and traceability pipelines. However, reliably localizing their dot structures remains challenging due to specular reflections, surface roughness, machining artifacts, and local damage. We propose a hybrid deepgeometric localization framework integrating YOLOv10-based dot detection with subpixel geometric refinement and projective grid reconstruction. First, YOLOv10 robustly detects dot candidates under severe illumination disturbances. A sub-pixel refinement module further enhances geometric precision by combining intensity-weighted centroid estimation and Taubin circle fitting. To suppress false detections arising from scratches or metallic noise, a geometric consistency filter based on inter-dot distance statistics and density clustering is applied. Subsequently, a RANSAC-based homography estimation aligns the detected dots with an ideal DM template for global grid reconstruction. Missing dots caused by reflections or damage are recovered using a reprojection-driven completion strategy. Experiments on real-world industrial metal datasets show that the proposed method significantly improves localization accuracy, grid completeness, and decoding robustness compared with deep-learning-only and geometry-only baselines.

• Publication year

  2025

• Venue

  2025 IEEE 4th Industrial Electronics Society Annual On-Line Conference (ONCON)

• Publication date

  2025-12-11

• Fields of study

  Not labeled

• Identifiers
  DOI 10.1109/ONCON68412.2025.11384492
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• A Data Matrix Code Recognition Method Based on L-Shaped Dashed Edge Localization Using Central Prior

  2024cited by this paper
• RFAConv-CBM-ViT: enhanced vision transformer for metal surface defect detection

  2024cited by this paper
• YOLOv10: Real-Time End-to-End Object Detection

  2024cited by this paper
• DM Code Key Point Detection Algorithm Based on CenterNet

  2023cited by this paper
• Data Extraction Method for Industrial Data Matrix Codes Based on Local Adjacent Modules Structure

  2022cited by this paper
• A New Outlier Removal Strategy Based on Reliability of Correspondence Graph for Fast Point Cloud Registration

  2022cited by this paper
• Two-Dimensional Codes Recognition Algorithm Based on Yolov5

  2022cited by this paper
• A Multi-Task Network for Joint Specular Highlight Detection and Removal

  2021cited by this paper
• Comparative Study of Data Matrix Codes Localization and Recognition Methods

  2021cited by this paper
• The realization of identification method for DataMatrix code

  2017cited by this paper
• AprilTag 2: Efficient and robust fiducial detection

  2016cited by this paper
• Fast Algorithm for Recognition of 2D Barcode: A Review

  2016cited by this paper
• Data Matrix Code Location Based on Finder Pattern Detection and Bar Code Border Fitting

  2012cited by this paper
• Robust and Fast Localization Algorithm for Data Matrix Barcode

  2010cited by this paper
• Polarization-Based Material Classification from Specular Reflection

  1990cited by this paper
• Surface Reflection: Physical and Geometrical Perspectives

  1989cited by this paper
• A Reflectance Model for Computer Graphics

  1982cited by this paper
• A reflectance model for computer graphics

  1981cited by this paper

• No citing papers are available for this paper.