Inspection and feature specification of the fancy yarns using diffraction limited incoherent imaging

Abstract

Chenille yarn features (pile yarn length, twist, hairiness, defect length, coefficient of diameter variation) were determined by diffraction limited incoherent imaging (DLIM) allowing to simulate sensor plane irradiance. The predicted image irradiance used in determining these features were obtained by deconvolving the captured image irradiance with incoherent point spread function designated by rectangular aperture. In twist values obtained by proposed DLIM, mean absolute relative error (|RE|mean) was determined as 0.822% by comparing Uster Twist Tester results. The DLIM in determining diameter provided better results with 0.8194% |RE|mean, compared to that 2.044% of optoelectronic method. The determination of hairiness fibers exceeding the ordinary lengths of pile yarns using Fourier domain high pass filtering (HPF) whose size is designated by the cut-off frequencies is proposed. These cut-off frequencies were determined in response to mean diameters obtained by image processing. The results of hairiness indices of HPF were compared with image processing results.