Study On Measuring Microfiber Diameter In Melt-blown Web Based On Image Processing

As a high quality of technical textile material, superfine fiber has excellent filtration performance because of its small diameter, large surface area, high pore ratio, and strong absorbility. Melt-blown nonwovens are widely used in the filter material, and fiber diameter plays an important role in affecting the filtration efficiency of melt-blown web. The traditional measurement is carried out manually to measure fiber diameter in SEM, thus, its efficiency is low, the labor intensity is huge, and the accuracy may decrease as the operators are getting tired. It is a strong requirement for seeking for a accurate measurement, it is also the purpose of our research to propose a solution of measuring the fiber diameters quickly, automatically and accurately.This paper studies on the measurement of superfine fiber diameter in melt-blown material using images gained from optical microscope and scanning electron microscope. A large number of fiber images of different fabric weight are captured in optical microscopy. It's found that the images captured are vague because of the serious multi-focal plane phenomenon caused by the small fiber diameters and the thick material. These images are further processed with Otsu's global threshold, LOG and CANNY edge detection operator, expected result still have not achieved. The experiment comes to conclusion that the idea of automatic analysis using optical microscope images was difficult to achieve. For SEM images, it proposes one set of image processing algorithms, which will be used in processing the SEM images. Based on edge detection, it uses image segmentation method of iterative threshold value to differentiate the goal fiber from image background, and the fiber edge line skeleton is obtained after the refining step. Utilizing line segment extraction and connection, the edge line is ready for rectangle pairing and finally the fiber diameter is gained by calculating the width of the rectangle.This paper has designed two sampling methods, the entire and the regional sampling, with the latter including 4 kinds of different sampling plans. Using these sampling methods, the SEM images of 12 melt-blown materials are processed, and the fibers diameters are gained. In the meantime, fibers diameters are measured manually using specific measurement software, and the diameters data are analyzed about its accuracy and stability. These data gained by manual measurement is used as standard data to evaluate the image processing result. This paper also introduces the tests about the validity, reproducibility and stability of the whole set of algorithm. The test result has indicated that this it is good at processing the SEM image of melt-blown material, and the data obtained is accurate. The algorithms itself has achieved certain validity, reproducibility and stability.In addition, this paper discusses the remaining problems, analyzing the cause reason and the possible solution. After summarizing the performance of this set of algorithms, it proposes one new set of possible algorithms which might be useful for the improvement of the project.