Applying wavelets transform, rough set theory and support vector machine for copper clad laminate defects classification

In this paper, we present a multi-resolution approach for the inspection of local defects embedded in homogeneous copper clad laminate (CCL) surfaces. The proposed method does not just rely on the extraction of local textural features in a spatial basis. It is based mainly on reconstructed images using the wavelet transform and inverse wavelet transform on the smooth subimage and detail subimages by properly selecting the adequate wavelet bases as well as the number of decomposition levels. The restored image will remove regular, repetitive texture patterns and enhance only local anomalies. Based on these local anomalies, feature extraction methods can then be used to discriminate between the defective regions and homogeneous regions in the restored image. Rough set feature selection algorithms are employed to select the feature. Rough set theory can deal with vagueness and uncertainties in image analysis, and can efficiently reduce the dimensionality of the feature space. Real samples with four classes of defects have been classified using the novel multi-classifier, namely, support vector machine. Effects of different sampling approach, kernel functions, and parameter settings used for SVM classification are thoroughly evaluated and discussed. The experimental results were also compared with the error back-propagation neural network classifier to demonstrate the efficacy of the proposed method.