Process knowledge based multi-class support vector classification (PK-MSVM) approach for surface defects in hot rolling

Article ID	Journal	Published Year	Pages	File Type
10323009	Expert Systems with Applications	2011	12 Pages	PDF

Abstract

âº We investigate the random surface defects on hot rolled steel bars and coils. âº Automatic image based inspection technique used for investigation. âº Defects classified based on process knowledge and image processing. âº Process knowledge leads to better classification scheme by support vector machines.

Keywords

Automated inspection Process knowledge Classification Support vector machines Hot bar rolling Image processing

Related Topics

Physical Sciences and Engineering Computer Science Artificial Intelligence

Preview

Process knowledge based multi-class support vector classification (PK-MSVM) approach for surface defects in hot rolling

Authors

Kuldeep Agarwal, Rajiv Shivpuri, Yijun Zhu, Tzyy-Shuh Chang, Howard Huang,