Key Method Research And Design Of Slubby Yarn Outward Parameters Detection

Textiles made of slubby yarn are featured by outstanding figure, uniquestyle, and strong stereoscopic effect. Thus, slubby yarn is widely used in textileindustry. Three parameters including bamboo joint length, bamboo distance andbamnoo ratio of slubby yarn need to be detected due to its unique outwardappearance, which is harder than detecting ordinary cotton yarn.Traditionally,there are two detection methods.One is man-made method primarily representedby blackboard detection method, which wastes a lot of time and energy andintroduces error casued by subjective factor and a small amount of samples.Theother is indirect measure method primarily represented by Uster capacitancemethod, which wastes a little time and energy and works automatively. However,this method also brings error becasure of introdution of physical quantity andthis method costs a lot.We need to acquire high-quality slubby yarn image before slubby yarnoutward parameters are detected. Thus, this paper has designed a suit of slubbyyarn image acqusition system in order to acquire high-quality slubby yarn image.It wastes a big amount of memory to acquire slubby yarn image directly becauselarge-scale image data need to be acquired.In the process of acquiring image, thus, this paper adopts double-buffering technology to gather and store slubbyyarn image so that memory can be saved substantially. Traditional methods willfail to display large-scale image data when image need to be displayed. Thispaper comes up with a method that allocates a small memory to displaylagre-scale image indirectly.This paper need to acquire diameter series of slubby yarn via the method ofimage processing in order to make a distinction between bamboo joint andbamboo distance. Slubby yarn image should be segmented before diameterseries are extracted.This paper adopts2-D maximum entropy genetic algorithmto segment Slubby yarn image in order to enhance robustness and efficiency ofimage segmentation. This paper makes use of open operation from themorphology to process slubby yarn image so that image hairiness can beremoved.This paper brings forward a kind of sorted series clustering algorithmto classify acquired diameter series. For the purpose of detection accuracy, akind of window filtering method will be designed to optimize classified result.This paper adopts object-oriented software engineering method to design thestruture of software, which enhances extension and reusability of software.