Appearance Quality Detection Of Nonwovens Based On Image Processing

Non-woven fabric is a fiber aggregate, the fiber material properties and the propertiesof the nonwoven fabric is closely related. Diameter of the fiber affects the friction propertiesof the nonwoven fabric, strength, bulk density and texture; The nonwoven fabric filterperformance is depended on the pore size distribution; Defects that reflects the non-wovenfabric mechanical properties, and it is also an integral part of evaluation indicators. Byanalyzing the non-woven fabric appearance quality parameters, we can grasp its generalperformance. While we have the general method for measuring parameters of the nonwovenfabric’s appearance, but there are certain limitations and disadvantages. Using imageprocessing techniques for measuring the non-woven fabric fiber diameter, orientationdistribution, and pore size distribution and detection of defects then compared to traditionalmethods. In this paper, we studied on the non-woven fabric appearance parameters by usingthe image processing, after obtaining the sample image, we will test each of the threeparameters. For the measurement of the fiber diameter, the fiber obtain the edge linesaccording to the edge detection, by using line by line or column by column scanning,measured positions of the two edges of the fiber through curve fitting method proposed edgeline together, the slope of the two edges will be obtained, then inclination of the fibers will beobtained, and then we can obtain the diameter of the fibers according to horizontal orvertical geometry of fiber width and fiber angle, Since the boundary line of the web holes inthe vicinity of both represent the inclination of the fibers, so that the holes can be calculatedby the edge line to replace the fiber inclination angle, thereby to obtain the fiber’s orientationdistribution, Because the area of the holes is too small, we can view them as a point, sowhen the extraction hole image, we need to set a critical area as a reference, Bycharacterizing morphology of the nonwoven web parameters, we can obtain the non-wovenpore size distribution. The parameters include pore fractal dimension, porosity, number ofholes, hole size, these parameters can reflect the size and distribution of pore.Impurities, holes and oil classes defects are the most common defects, Therefore, thispaper study and test these three defects, we detect the defects by using the principle whichthe different defects have the totally different effect of different types of reflection andtransmission of the light.