On-line yarn evenness detection using CCD image sensor

This paper introduces the method of on-line measurement of the yarn evenness using a CCD image sensor. The sensor is mounted on a movable bracket of the small car. The obtained yarn coefficients of variation in the diameter through computer and digital image processing techniques are compared with those of using off-line USTERII evenness tester. To a certain extent, avoiding affection of the incident light intensity, the CCD sensor overcomes the shortcomings of the conventional photoelectric sensors. The sensor also uses a 250,000-pixel CCD equivalent to that of a conventional machine vision sensor. This allows presence detection and recognition of different objects, which have up till now been performed by a human, to be executed fast and accurately. Combined with modern control theory, fuzzy control and expert system are designed to collect and process real-time data of the measuring device. The results of on-line measurement show that the average error accuracy is less than 5.5% in the study. In addition, it is also found that the achieved data from the on-line device are coherent and correlative with USTER statistical bulletin 2001. Therefore, the development of the portable, higher cost-effective measuring device has a positive role in promoting the textile industry and improving the quality of the yarn.