comparison circular knitted fabric defects detection and classification using machine vision

this paper describes a computer vision aided fabric inspection system to detect and classity circular knitted fabric defects using common different texture recognition methods including co -occurrence matrices discrete fourier transform, wavelet gabor and clustering the fabric images were broadly classified into six classes cracks, holes vertical stripes horizontal stripes soil freckle and defect free 120 images 256 gray level and 100 dpi containing 20 images o fdefect free fabrics rib1×1 and also 20 images correspoding to five different categories were used. in general one -half images of each category employed for training and remaining images were used for testing.

کلیدواژه ها:

computer vision ، fabric inspection ، circular knitting machine ، neural network ، co-occurrence matrix ، fourier transfrm wavelet transfiorm gabor transform and clustering