Surface Defect Detection Using Texture Features and RBFN
This paper is concerned with the problem of automatic inspection of hot-rolled plate surface using machine vision. An intelligent surface defect detection paradigm based on texture analysis and neural network is presented. Texture features based on GLCM, Laws energy, and LBP are extracted from ROI in hot-rolled plate surface images. These features are integrated into a feature vector which uniquely differentiates the abnormal regions from normal surface. A radial basis function network is used for classification of ROI as normal or abnormal. Classification accuracies using the individual feature sets and the integrated features are compared. The results indicate that the integrated features improve the accuracy of detection. Empirical results show the integrated features from GLCM and LBP perform well in classifying the samples with the lowest classification error.
Q. Song "Surface Defect Detection Using Texture Features and RBFN", Applied Mechanics and Materials, Vols. 44-47, pp. 3529-3533, 2011