Abstract:

In order to more fully characterize the surface information of the fabric to facilitate the subsequent fabric structure analysis, appearance evaluation, quality inspection and reverse engineering and other applications, a three-dimensional contour reconstruction method and system were developed for realizing the digital image acquisition, camera calibration and three-dimensional reconstruction. A multi-angle image acquisition platform for textileswas set, the influence of the imaging parameters optimization and the algorithm optimization on the three-dimensional reconatruction effect was investigated, and the experimental process and shooting environment was gradually optimized to improve the system accuracy of the three-dimensional reconstruction of textiles. The results show that based on existing hardware, the platform can complete high-precision reconstruction of the three-dimensional contour of low density textiles, and the reconstructed model can carry a large number of texture information, spatial position relation and color information on the surface of the fabric, and the feature point number is above 1 000. 

Key words: textile testing, camera calibration, feature point matching, point colud fusion, three-dimensional reconstruction