A fast and efficient algorithm is proposed to detect and extract geometric features of a fabric wrinkle. Its boundary attributes may be extracted firstly by detecting and analyzing its 2-D image. After extraction, such geometric features are possibly represented by the 1/f noise curve. According to the different light intensities in the wrinkled area, the wrinkle's height values can be achieved and then some useful curve formulations are generated. Therefore, the curved surface is finally shaped by adopting a curve-fitting technique. As a non-iterative algorithm is employed among detection and extraction processes, the close correlation between the predicted and the experimental results seems to prove that the fabric wrinkles' 3-D geometric features may be detected and extracted fastly and efficiently.
展开▼
