Research On The Automatic Detection Method Of Fabric Density Based On Image Processing

In the country’s greige cloth standard, the yarn density is one of the important indicators tomeasure the quality of the fabric. In the current textile enterprises, the detection of yarn densityis relied on the professional tools such as the cloth prover. This manual detection method has alow degree of automation and efficiency and the labor intensity is too large for workers, so it cannot guarantee the detection quality. With the increasing development of various varieties andsmall bath production of mode in the textile enterprises, we need to research the method ofautomatic detection and form the relates products. It will be conducive to simplify the process ofdetection and improve the efficiency of detection. It will also has an important significance toenhance the quality and value-added of products and promote the development of the industry.For this purpose,two method of automatic detection of woven fabric density are presented inthis paper. The key is that we can design the corresponding algorithms to improve the efficiencyand reliability of woven fabric density detection through processing and analyzing the image ofwoven fabric. In the actual process of study, in order to improve the precision of the detection ofwoven fabric, this paper achieves a algorithm for tilt the image of woven fabric which is base onthe Hough transform and achieves a rapid image correction by image rotation. It can resolve theproblem that the woven is difficult to maintain a horizontal and vertical location. In addition, thispaper gives two algorithms which is based on Fourier transform and wavelet transform in orderto detect the woven fabric density automatically. We can get a power spectrum of woven fabricimage by Fourier transform so that it’s convenient to us to detect the warp density and weftdensity automatically by the relationship of the spatial domain and frequency domain. Moreover,we apply the wavelet transform to obtain the warp and weft sub-image of the image of wovenimage, then we can get the desired information of warp density and weft density by binarizingand smoothing the sub-image and design the procedures to calculate the woven fabric densityautomatically. Furthermore, this paper uses lots of experiment data to verify the reliability andapplicability of the Fourier transform and wavelet transform utilized to detect the woven fabricdensity from several angles, such as the spacing level of fabric, different discriminating regions.In the paper, we use the image processing technology and textile expertise to detect thedensity of woven fabric. We propose several algorithms to solve some key problems in this process. This study has great theoretical significance and application value for the detection ofwoven fabric density and development of the textile industry.