Research On On-line Detection Technology For Zipper Dimension Based On Machine Vision

Nylon zipper is an important accessory of clothing and luggage.Due to the artificial,equipment and other reasons,there are many defects such as chain tooth size unqualified.Those defects seriously affect the performance and the quality of zipper.At the same time,the traditional manual testing method has the lag and instability of results,low efficiency and high cost,which has seriously restricted the intelligent and informational development of the enterprise.This paper applies the machine vision to the on-line detection of nylon zipper dimension,and designs a set of on-line visual detection system for nylon zipper dimension,including zipper image acquisition device,the algorithm of zipper dimension detection and vibration compensation.The main contents and conclusions are as follows:1)Design of image acquisition scheme of the on-line visual detection system for nylon zipper dimension.According to the requirements of nylon zipper chain teeth containing multiple appearance size detection objects(such as single thickness,single width and tooth head),the key hardware of the image acquisition device was analyzed and selected,and the corresponding image acquisition optical path was designed.It is used to simultaneously collect two single side zippers produced by the zipper forming machine are and to collect images of different positions of the single side zipper step by step through a single industrial camera.2)Research on the size detection algorithm of nylon zipper Firstly,the distortion correction of the zipper image was realized through the camera calibration principle,and the pixel equivalent of the corresponding image was calculated.Then,the image preprocessing methods of zipper images were studied,including ROI extraction,image filtering and denoising,and threshold segmentation,etc.,thus to reduce the interference information in the image and highlight the zipper being measured.After that,according to the detection object of different appearance size of nylon zipper,the corresponding edge size detection algorithm is proposed to realize the actual size detection of the zipper.The experimental results show that: the uncertainty of the measurement results of single thickness,single width and tooth head of nylon zipper were 0.00180 mm,0.00084 mm and 0.00084 mm,less than 0.002 mm of standard request;and the error rate in the online detection was 1.36%,1.36% and 1.53% respectively,which was less than the standard error rate of 2%.It is proved that the detection algorithm of zipper dimension proposed in this paper meets the requirements of accuracy and reliability of zipper dimension measurement.However,the error analysis indicated that the influence of vibration on the measurement results should be further studied.3)Research on vibration compensation algorithm in nylon zipper dimension measurement.Firstly,this paper analyzed of the zipper vibration mechanism,and determined the vibration of production equipment has a great influence on the detection results of single thickness zipper.Then,the differential equation of the zipper transverse vibration model was established based on the Hamilton variation principle,and solved the approximate analytical solution of the zipper transverse amplitude.After that,according to the measurement principle of LK optical flow method,a global vibration information calculation method based on the optical flow method was proposed which used to calculate the transverse amplitude of the zipper at a certain time.Finally,an improved detection algorithm of nylon zipper single thickness dimension based on vibration compensation was proposed.The experimental results show that: the uncertainty of the single-thickness zipper measured by the improved algorithm was0.00051 mm,which was 71.7% less than the original detection algorithm;and the error rate of online detection by the improved algorithm was 0.61%,which was 58.5% lower than the original algorithm.It is proved that the improved algorithm based on vibration compensation has further improved the detection performance compared with the original algorithm.The above work provided methods and technology foundation for the on-line detection of nylon zipper dimension based on machine vision,which contributed to the automation of the detection of nylon zipper dimension.