The Design And Research Of The Fiber Automatic Recognition Offline Software System

This paper uses digital image processing and pattern recognition theories toinvent automatic identification through fiber images.Fiber testing, identifying the compositions and their respective share, is ofsignificant importance in assessing quality analysis performance of textiles. Itincludes the closure of the pipeline through the dust cage network connectivity,metering devices, opening devices, storage of cotton-channel, combing device, and acombination of electrical control systems. It bears the followingcharacteristics:setting up unusual fiber detecting device between Combing deviceand poly cotton plant, excluding foreign fibers device detection device described infiber, including transparent pipes, lighting fixtures, high-speed camera, andinstalling image recognition system equipment control circuit and control software inthe computer.Fiber image automatic identification system can automatically identify the fiberimage scene, uses micro-fiber optical fiber detector to obtain images, gets access tocomputer technology through the USB bus interface. This system primitivelypreprocess fiber images, through the median filter, smoothing filter and sharpeningtechniques, then gets the GLCM for fiber image, during which the energy, entropy,inertia, correlation and partial smoothness are extracted. Finally, use of nearestneighbor method for automatic classification of fiber types. The system distinguishitself from the existed progress that is complicated and subjective in manuallyidentifying textiles.The software for fiber image processing and analysis, programmed in Visual C++6.0, an visual programming language,has integrated fiber image processing andanalysis algorithms, including image preprocessing and image feature extractionalgorithms. Database management system for fibers image is embedded in thissoftware, thus extracted image features characteristic values can be directly storedin the database for further analysis or identification. Results of image processing and analysis can be displayed on the screen, which is user-friendly and easy to operate.Fiber detection software can be used offline, thus the inspectors greatly arefreed from handling the heavy devices,which is necessary for the large existeddetecting machines. Fiber-detection software stores the test results in the computerdatabase, and image processing system uses advanced image processing technologyto accurately measure or calculate the fiber diameter and density scales. Fiber imageautomatic recognition software improve the Level of automation for detecting,and isa good application that is economically beneficial.