Fabric Defect Detection System Based On Zedboard

The quality of fabric become low because of defects,many small and medium-sized enterprises want to import fabric defect detection platform.However,the high price,the limited type of defects and the high difficulty to develop platform make them not do it.So the enterprises have to rely on human to identify the fabric defects.Human defection,however,has shortcomings like visual fatigue and poor judgment accuracy,which fails to achieve the high efficiency the identification of fabric defects.In order to solve problems above,this paper studies the fabric defect detection platform to ensure the quality of the fabric while reduce the cost of production.The Zedboard developed by Xilinx company was adopted as defect detection platform in present study,which lie in exploring the following issues:(1)The subject in this study is the fabric defect.In terms of fabric defect detection method,the classical algorithm based on gray level co-occurrence matrix for texture to acquire feature values has low efficiency,which can not meet the real-time requirements of fabric defects.Therefore,this paper presents an improved gray level co-occurrence matrix,namely,category co-occurrence matrix,by the extraction of image texture features based on texture feature and set the threshold of fabric to finish defect detection.(2)Fabric defect progress are as follows:firstly,the pilot action of fabric image has been done,which included the fast median filtering,histogram equalization,and illumination compensation adjustment.Then,the noise of image has been eliminated so as to highlight the defect and improve the light effect.At the end,the image was segmented by adopting the co-occurrence matrix algorithm and the result gained was compared with the result based on the local entropy and energy.(3)Taking Zedboard as the development platform,the current project adopted the CCD camera and LCD screen as hardware platform for fabric defect detection.Zedboard contains FPGA and ARM.The design of defect detection algorithm and algorithm of IP core has achieved on the FPGA.Then,the controlling of the image acquisition has finished on ARM,and the driver configuration of IP kernel,the transmission of Qt so as to build real-time image display interface have completed in Linux system.Lastly,the OpenCV library was adopted to realise the fabric defect detection by software.(4)The author built the hardware platform to tell the defect detection algorithm.The experimental results show that the system has high detection efficiency and high precision,and can be used in textile enterprises.