<title>Optical processor for real-time detection of defects in textile webs</title>

<title>3D vision with 2D images by mediating the analysis of shade areas for the illumination angle with the surface</title>

“…This way the regular texture of the textile becomes visible [14] and possible shortcomings can be observed. A compound image of the previous shows the textile structure, isolating the finish component that defines the structure of the textile in a high level [5].. On these images, a recognition of each situation of the textile can be carried out by using different procedures; such as: methods of spectral analysis and filtrates [8], [7] by using techniques of correlation in the fourier dominion [9]; Gabor filters [10]; Moire Techniques [1 1]; Mensuration of profiles of density in the woofjourneys as well as of warp [6]; Extraction of characteristic in iterative maps [12]; as a recognition of patterns of each of these situations [13].…”

Contribution of polarimetric imaging for the characterization of fibrous surface properties at different scales

“…Wood [9] also uses the Fourier transform in order to analyze the modifications of the state of surface of textile carpets through the modifications of their texture properties. Several filtering techniques are used in order to process these signals and for instance Tsai [10, 11] Shakher [12] or Kreißl [13] use filtering methods based on wavelet transform. These processes allow the user to obtain images where only defects appear, without any periodical structure.…”

“…Whereas most authors classically take intensity snapshots they process with more or less complex image processing techniques [8][9][10][11][12][13][14][15][16][17][18][19][20][21], we rather intend to extract the relevant information with optical techniques before grabbing and implement subsequent data extraction from 2-D acquired signals. The latter data extraction often proves rather simple and therefore less subject to noise than classic image processing techniques reported in [8][9][10][11][12][13][14][15][16][17][18][19][20][21]. This approach can seem less flexible, since it requires the development of specific apparatus, but it allows us to extract information hardly visible in intensity images, which cannot be recovered by image post-processings [3,4,22].…”

Novel method for patterned fabric inspection using Bollinger bands

“…[1][2][3] There is no designed pattern on the fabric surface. Many methods such as autoregressive method, 4 neural network, [5][6][7][8][9] Fourier transform, [10][11][12][13] Gabor filters, [14][15][16][17][18] wavelet transform, [19][20][21][22][23][24][25][26] structural method, 27 and some spatial and spectral approaches 28 have been developed for them. Most detection success rates of these methods can reach over 95%.…”