The Research Of Woven Fabric Density Detection Algorithm Based On Image Processing

In the industry, woven fabric density has significant influence on the value of textile. Traditional density detection is based on the artificial detection, and in the entire processing of testing, it requires examiner spirit to remain highly concentrated. If there is any error in the process of testing, it will begin again. The intensity of labor of inspector is very larger, and the processing is consuming time and effort, of course, which will result in quite low efficiency, and is facilitate to damage the woven fabric. At the same time, the examiner is in a state of highly concentration and nervous for long, and the body health is affected largely. As the development of machine vision and image processing technology, it realizes the automatic detection of woven fabric density, which not only shortens the production process and improves the production efficiency, but speeds up the development of the textile industry automation production and quick implementation of intelligentialize.The content of this paper is mainly divided into five parts as follows: the first chapter is introduction. It is mainly the comparison for automatic detection of woven fabric density image processing researched by home and abroad to see the current shortcomings and disadvantages of the method for automatic detection for fabric density. Based on this, some researches for fabric density detection will be presented in this paper. The second chapter principally introduces fabric image pretreatment, skew detection and correction, which is on account of being inevitable that the captured images will be skew to a certain extent and contain noise during the capturing processing which will result in larger error for fabric density detection. In chapter 3 and chapter 4, two research methods: based on wavelet transform and based on the gray projection method for fabric density detection are described. Firstly, the preprocessing images are decomposition and reconstruction by wavelet transform to separate the warps form the wefts. Then morphology processing, such as binarization and smooth, is done for the images and the clear and smooth of image yarns are obtained, eventually the density of yarns are achieved again. Combining with the biorthogonal wavelet as a filter, gray projection curve is utilized for fabric density detection. Then the gray projection curves in vertical direction and horizontal direction respectively are drawn, which conclude many peaks and each peak stands for one yarns. Hence, according to the peaks, the yarns density could be achieved. As an extension of the woven fabric density detection and based on the combination of Lab VIEW and MATLAB, an interface of woven fabric density detection is designed in the last part of the main content.In the experiments, the woven fabric images preprocessing and skew detection and correction are implemented firstly and better enhanced fabric images are obtained; select suitable wavelet and realize the automatic detection for woven fabric density based on the wavelet transform gray level projection. Via MATLAB, it achieves the automatic calculation of woven fabric of density; finally, combing Lab VIEW and MATLAB, a relatively simple interface fabric density is achieved.On the basis of digital image processing, the methods for woven fabric density detection are proposed in this paper presents. It not only avoids the influence of subjective factors and makes the result objective and accurate, but improves the detection efficiency, reduces the artificial detection error, protects health inspector, and realizes high quality production and diversity of the quick design in textile, which is confirmed to be a good application prospect.