تصویربرداری و پردازش تصویر در تحقیقات رسانه متخلخل Imaging and image processing in porous media research

1. Introduction Several projects have used imaging methods and image analysis to describe and model the flow behavior in soil and more generally in porous media [1–6]. Non-destructive image acquisition such as computed tomography (CT) has the advantage over destructive imaging that the same sample can be scanned repeatedly under different initial conditions or a process can be monitored spatially and temporally. Traditionally, samples are exposed to X-rays that are sensitive to density variations in the sample. In our measurements, we combine this modality with illumination by thermal neutrons. The beam from this source type is sensitive to hydrogen atoms. By imaging modality we mean imaging systems that are based on different physical concepts and hence provide images that reveal different information. Thus, unlike many metals, water and organic matter strongly attenuate and scatter neutrons. Examples of applications using images are structure quantification [7–9], flow modeling [5,10], and mechanics of porous media [11,12]. The main image processing tasks are noise reduction, image enhancement, pixel classification, and pixel clustering. We review both image processing research regarding the development of processing methods and research of method applied directly in the field of porous media. The performance of some of the methods described will be demonstrated. Here, we use a three-dimensional image of a partially saturated sand sample. We detail the useful methods in our projects and combine the results in a multi-scale and -modality context. In Sections 2–4, we describe the imaging technologies we used, the image processing, and an example that shows how the combination of neutron and X-ray-based imaging, along with image processing can lead to new insights in porous media research. The interpretation and quantification of the images requires intensive image analysis.