DWT Based LCP Features for the Classification of Steel Surface Defects in SEM Images with KNN Classifier
Year: 2016
Type: preprint
Abstract: This paper presents a Texture based approach for detecting and characterizing defects on steel surfaces, through which the quality inspection of steel products is performed. Our objective is to detect the defects in the surface of steel products. Major defects like corrosion, crack, scratch and fracture are considered for testing and training process. The Discrete Wavelet Transform (DWT) based Local Configuration Pattern (LCP) features are given as input to the KNN classifier. In order to explore multichannel discriminative information of both the microscopic configuration and local structures, DWT based LCP which is a combination of Local Binary Pattern (LBP) and Microscopic Configuration modeling (MiC) method is proposed. LCP and Defect Severity Index values are calculated for the sub-images obtained after applying DWT. The overall accuracy of DWT based LCP method is 96.7%. The results show that the proposed method produces better classification accuracy when compared to the classical methods.
Source: SSRN Electronic Journal
Authors: S. Karthikeyan, M Pravin, B. Sathya Bama, M. Mareeswari
Institution: Thiagarajar College of Engineering
Language: English
Subfield: Industrial and Manufacturing Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3137139253
