Analysis On The Surface Structure Of ACF Micro-Image

ACF (Activated Carbon Fibers, ACF), also known as fibrous activated carbon, is one high-performance active adsorption material with a better performance than ordinary activated carbon (Activated Carbon, AC) as well as an excellent environmental engineering materials.ACF material images studied in this paper are obtained by SEM (Scanning Electron Microscope), so there are many features such as uneven illumination and low-contrast of foreground and background to this kind of images. In this essay, ACF material images are further researched and analyzed from four aspects:image preprocessing, edge detection of objects, segmentation of holes in interested areas and extraction of parameters.Traditional methods of image preprocessing are firstly discussed in this paper. For the feature of ACF material images, filtering algorithm, edge thinning operator, image sharpening operator and other enhancement algorithms are used to lay the foundation for subsequent segmentation and edge detecting. Based on the result of the image preprocessing, active contour model is adopted to detect edges of objects and its results are compared with curve fitting. Then, parameters of pores are calculated. A new method of edge tracking named as tangent direction method is also introduced in this article. This way aims to solve the problem that there exits significant grayscale mutation in the edge of the object. After accurate cut-off points are found, edge linking and roundness calculation are proposed to obtain the distribution of pores. For those images of which results got by active contour model and tangent direction method are unsatisfied, circle fitting combined with multi-threshold segmentation and Hough Transform is proposed. Finally, to deal with the holes adhesion, different watershed algorithms (watershed based on distance transform, watershed based on image gradient and water based on control marker) are used to segment images.