|About this Abstract
||2019 TMS Annual Meeting & Exhibition
||Aluminum Alloys, Processing and Characterization
||D-61: Improvements for The Recognitionn Rate of Surface Defects of Aluminum Strips
||Xiaoming Liu, Ke Xu, Dongdong Zhou
|On-Site Speaker (Planned)
Surface defects have significate effect on the mechanical properties of aluminum strips. However, the current defect identification algorithm needs to be further improved to meet the requirements of aluminum quality inspection. Algorithms of non-subsampled shearlet transform and the kernel spectral regression (NSST-KSR) are applied to identification surface defects of aluminum strips. The defect images collected from a cold rolling line of aluminum srtips were tested, including five true defect types of point imprints, scratches, dents, roll marks and wrinkles, and three pseudo defect types of lighting variation, water marks and oil stains. The results show that the overall recognition rate reached 95.83%, which is higher than the existing method in the literature. NSST-KSR method can effectively and steadily improve the recognition rate of aluminum strips surface defects.
||Planned: Light Metals