Automatic Defect Detection of Yarn-Dyed Fabrics Based on Energy Fusion and Local Binary Patterns

Wen Yu Li; Long Di Cheng; Wen Liang Xue

doi:10.4028/www.scientific.net/AMR.472-475.3039

Paper Titles

Process Modeling of Virtual Enterprises Partner Selection Based on Petri Net
p.3017

Research Review of Relationship between TQM and Organization Innovation
p.3023

Positioning Error Compensation for Machining Center Based on Spline Interpolation
p.3029

Forecast Research of Diamond Grinding Parameters Based on Multi-Variable Grey Model
p.3035

Automatic Defect Detection of Yarn-Dyed Fabrics Based on Energy Fusion and Local Binary Patterns
p.3039

Computer Control System Development for the New Type Pilot Cold Rolling Mill
p.3043

Online Estimation System of Grinding Wheel Status Based on D-S Evidence Theory
p.3047

Research of Grinding Wheel Wear Modeling Based on Hypothesis Test Principle
p.3053

Control and Optimizing of Unstable State in Continuous Casting Process with Automatic System
p.3057

HomeAdvanced Materials ResearchAdvanced Materials Research Vols. 472-475Automatic Defect Detection of Yarn-Dyed Fabrics...

Automatic Defect Detection of Yarn-Dyed Fabrics Based on Energy Fusion and Local Binary Patterns

Abstract:

For the purpose of realizing fast and effective detection of defects of yarn-dyed fabric, and in consideration of the inherent characteristics of texture, i.e., color and structure, an approach for automatic defect detection is proposed in this paper. The image of yarn-dyed fabric to be enhanced is first converted from RGB true color space to L*a*b* color space. Then Log-gabor filters filter chromatic and brightness channels, and energy feature images are acquired after energy is fused between chromatic and brightness. Finally defects of yarn-dyed fabrics can be detected on the energy feature images using local binary patterns. The proposed method can detect colored and structural flaws. Experimental results for the defect detection from six kinds of yarn-dyed fabrics indicate that a high detection rate is achieved for the proposed method. It is fast enough to be possible for real-time application.