The Structure Design And Arithmetic Study Of Detecting System Of Viscose Filament Yarn

One of the most important reasons making Viscose Filament yarn unstable is the short of number of the Viscose Filament yarn according to the asymmetry of the cloth coloration.The traditional quality control of Viscose Filament yarn mainly relies on manual work, and it is too slow to reach real-time surveillance.As a new-style testing technic, machine vision testing system has the excellences of faster speed, higher precision, and non-contact testing available. It does well in conquering deficiency on manual work and improving both automatization and efficiency if it is used in the field of chemic-fiber quality control.The Detecting System of Viscose Filament yarn consists of Photo- electronic Imaging System, Circuit Controlling System.Interactive System. Aiming at the testing needs of viscose filament yarn, this paper particularly researches the design of the software and hardware of the Detecting System. And image processing calculating arithmetic.This system receives the image data through the photo-electronic imaging system.After learning ecumenic image processing methods, the particularity of viscose filament yarn image was analysed. According to the characteristic of viscose filament.we study the arithmetic of enhancement, segmentation, and feature extraction for the image of viscose filament yarn. Combine with Morphogical analysis.the count problem of viscose filament binary image was studied.A kind of arithmetic named single-pixel tracking for solving the problem of crossover of viscose filament was put forward.The experiment outcome based on a large number of images proved validity of these arithmetic well. The error of test is controlled within 3%.