Haar wavelet-based approach to locating defects in texture images

Abstract

In this paper, a novel Haar wavelet-based approach to extracting and locating surface defects in grey-level texture images is proposed. This new approach explores space localization properties of the discrete Haar wavelet transform (HT), performs task-oriented statistical analysis of well-defined non-intersecting subsets of HT spectral coefficients, and generates parameterized defect detection criteria with an installed additional capability to locate surface defects in defective texture images. The preliminary experimental analysis results demonstrating the use of the developed approach to the automated visual inspection of ceramic tiles, obtained from the real factory environment, are also presented.