Using Digital Image Processing Technology To Measure The Structural Parameters Of The Knitted Fabric

Digital Image Processing Technique is a kind of important tool that applied mathematics and computer to resolve actual problems, having reached to each realm. In recent years, with seeped into many fields about textile inspection and automatic recognition, it changes traditional manual operation in the textile industry, highly increasing work efficiency and solving many difficulties. This also predicts that it is a trend that applying Digital Image Processing Technique to resolve actual problems.This dissertation has applied Digital Image Processing Techniques to achieved knitted fabric's parameters that involve in the density, the stitch length, superficial hole's percentage and underfilled coefficient, etc. According to the characteristic of knitted fabric image, suitable algorithms such as media filtering, 2D Fourier transform, picking up characteristics in the frequency image, image reconstruct, histogram that can be partitioned by a single thresholding, thinning, etc, are put forward, which basically solved the problem of extracting useful signals in the fabric images. Induce transform formula between 2D and 3D of stitch's geometry model. After doing many experiments, we certified that the two methods have no obvious differences between automatic measurement and manual measurement.Under the Chinese Windows environment, using the Visual Basic 6.0 development tool, investigated Automatic Recognition System for Knitted Fabric Parameters. The interface is friendly, and operation is simple.This dissertation just studied the periodically knitted fabric, especially jersey pattern. It concludes that the automatic method can be adopted in the production, but we still need do more and more jobs to improve it for the application.