Texture Defect Detection Using Dual-Tree Complex Wavelet Reconstruction

Abstract

This paper introduces a new approach for automated inspection of textured materials using Dual-Tree Complex Wavelet (DT-CWT). The DT-CWT can transform images into a representation with six directionally selective sub bands for each scale. By properly selecting the smooth sub image or the combination of detail sub images in different resolution levels for backward wavelet transform, the reconstructed image will remove regular, repetitive texture patterns and enhance only local anomalies. The difficult defect detection problem in complicated textured images is converted into a simple thresholding problem in nontextured images. The experimental results show that the DT-CWT is more effective than the real discrete wavelet transform.