Online Surface Defect Identification of Cold Rolled Strips Based on Local Binary Pattern and Extreme Learning Machine

Abstract

In the production of cold-rolled strip, the strip surface may suffer from various defects which need to be detected and identified using an online inspection system. The system is equipped with high-speed and high-resolution cameras to acquire images from the moving strip surface. Features are then extracted from the images and are used as inputs of a pre-trained classifier to identify the type of defect. New types of defect often appear in production. At this point the pre-trained classifier needs to be quickly retrained and deployed in seconds to meet the requirement of the online identification of all defects in the environment of a continuous production line. Therefore, the method for extracting the image features and the training for the classification model should be automated and fast enough, normally within seconds. This paper presents our findings in investigating the computational and classification performance of various feature extraction methods and classification models for the strip surface defect identification. The methods include Scale Invariant Feature Transform (SIFT), Speeded Up Robust Features (SURF) and Local Binary Patterns (LBP). The classifiers we have assessed include Back Propagation (BP) neural network, Support Vector Machine (SVM) and Extreme Learning Machine (ELM). By comparing various combinations of different feature extraction and classification methods, our experiments show that the hybrid method of LBP for feature extraction and ELM for defect classification results in less training and identification time with higher classification accuracy, which satisfied online real-time identification.

Automated Visual Defect Detection for Flat Steel Surface: A Survey

“…As a classical operator, local binary pattern (LBP) is widely used to characterize local texture features of images, which has significant advantages of rotation and gray invariance. In 1994, LBP is first proposed by Ojala et al in [42], Later, LBP is frequently used to detect defects on flat steel surface [43][44][45]. In order to overcome the shortcomings of the original LBP (i.e., weak global descriptive and noise-sensitive), various LBP variants are developed based on changing the threshold or scale of the original LBP (refer to Fig.…”

Automated Visual Defect Classification for Flat Steel Surface: A Survey

“…LBP is one of the most successful local texture feature operators, which creates an intensity-and rotation-invariant binary descriptor and estimates the local contrast of an image based on the differences between adjacent pixels and central pixel, whose encoding mode and sampling rules are briefly given in Fig. 4, and it has been widely used to extract features of steel surfaces [11,[81][82][83][84][85][86][87]. In addition, some variants based on the original LBP have been proposed to overcome the limitations of LBP, such as noise sensitivity.…”

A Generic Adaptive Fractal Filtering Algorithm for Identifying Work Piece Defects on Multiple Surfaces

“…B.U. Heerdun and Feng Han [14] process the signal without dividing it into frames. According to their settings, the input signal of the filter should be a combined signal of sine wave and random noise,…”

“…However, it is still a challenge to inspect the defects using an automatic imaging system owing to miscellaneous patterns of the defects, low contrast between the defect and background, the existence of pseudo defects, as well as tiny workpiece surface defect detection problem such as small cracks and scratches are not easily identified by the conventional algorithms [9][10][11][12]. For surface defect identification application, Ke X. etal propose the methods include Scale Invariant Feature Transform (SIFT), Speeded Up Robust Features (SURF) and Local Binary Patterns (LBP) [13] and a novel nonsymmetry and anti-packing model (NAM) was proposed in [14].…”