Grade Assessment of Fabric Surface Wrinkling by Using Image Analysis

Abstract

Grade assessment of fabric surface wrinkling is a very important issue for assuring the quality of fabric seaming, which is inherently prone to deformation during garment manufacture. Although fabric properties can be related to seam pucker, the aesthetic judgement is still based upon ranking the stitched fabric samples by evaluators having compared them with photographic “standards”. Thus, the visual evaluations obtained are unreliable and time-inefficient, and do not always agree – the evaluators must therefore be trained. This paper reports the development of a new automotive measuring system which is based on the use of computer-aided vision, according to image information from a new image capture instrument which consists of a pair of laser sources and a webcam within a closed booth. In order to extract pucker information, the image of fabric seam first passes through a filter algorithm, then uses modified pixel reading to determine the exact coordinate of every point on seam curves. This measurement procedure is objective and is focused on the statistical analysis of pucker characteristics and distribution of seams of different grade levels, by modelling the cognitive process involved in seam pucker assessment. Based on this principle, a set of convenient computer interfaces were developed for operating grade assessment available in a company. The result of grade assessment from the system was consistent with experimental assessment from skilled engineers, which confirmed the suitability of the system for industrial applications.