Fabric defect detection based on adaptive local binary patterns
ROBIO'09 Proceedings of the 2009 international conference on Robotics and biomimetics
| Hi-index | 0.00 |
According to the property of wavelet transform and fabric texture's Fourier spectrum, a new method for defect detection was presented. The proposed method is based on wavelet lifting transform with one resolution level. By using restoration scheme of the Fourier transform, the normal fabric textures of smooth sub-image in the spatial domain are removed by detecting the high-energy frequency components of sub-image in the Fourier domain, setting them to zero using frequency-domain filter, and back-transforming to a spatial domain sub-image. Then, the smooth and detail sub-images are segmented into many sub-windows, in which standard deviation are calculated as extracted features. The extracted features are compared with normal sub-window's features to determine whether there exists defect. Experimental results show that this method is validity and feasibility.