Study And Design For Fabric Detect

In fabric production, quality control is important, which is ensured by fabric defectdetection. Nowadays, the fabric defect detection is still done by artificial vision inmajority of textile enterprises. This method is slower, higher false detection andundetected rate. Therefore, machine vision has become a hot topic at home and abroad.This study aims to build cord fabric defect detecting platform based on machinevision, at the same time to explore the algorithm for rapid and accurate cord fabric defectdetecting. Cord fabric has special texture, and moving speedily. Therefore it is difficult tocollect clear image, which is the key in the system.The hardware design of the system can be divided into five parts: CCD linear arraycamera, optical lens, image acquisition card, light source, computer and length measuringtool. CCD linear array camera and optical lens make up optical imaging system, lightsource for grabbing clear image provide enough intensity of illumination of imagesurface to make sure the image quality. Image acquisition card and computer make upimage grabbed part, and the image information grabbed from CCD linear array camera isstored in the computer. The length measuring tool can calculate the cord fabric runninglength, and provides the external synchronizing signal to the camera.According to the system requirements, this article chooses the hardware, and writesimage acquisition program using MIL8.0at VC++platform. This image acquisitionsystem has been completed installation and debugging. The results show that the systemcan collect clear cord fabric image in the industrial field.In the rapid fabric defect detecting field, this article describes the method of cordfabric defect detecting based on gray level histogram and the method based on gray levelprofile, meanwhile put forward the method based on ratio of gray cord fabric defect detecting. Gray histogram method is the basic one, but the false detection rate is higherbecause of its weakness. The method, based on gray profile coming from geomorphicprofile, is vulnerable to outside interference, that is, the method poor has poor stability.Ratio of gray is put forward based on texture feature of cord fabric, select the ratio of thebackground and foreground as the standard of the defect detection. This flexible methodis less disturbed by outside interference, and little contact with types of cord fabric.Therefore, compared with the first two methods, the ratio of gray method is more suitablefor rapid defect detection of cord fabric.