A protocol for obtaining digital images from natural porous media with a wide range of pore sizes, intended for fractal studies of the porosity, is proposed. Soil porosity is used as para- digm of complex natural poro...A protocol for obtaining digital images from natural porous media with a wide range of pore sizes, intended for fractal studies of the porosity, is proposed. Soil porosity is used as para- digm of complex natural porous media in this study. The use of several imaging devices and fluores- cent compounds to enhance the contrast between the solid and the pore phase is tested. Finally a protocol is reached using a photo camera and a confocal microscope. It is the first time that confocal microscopy is used for this purpose. Artificial porous images are created through random Sierpinski carpet fractals and the statistical information of real soil images. These ground truth images are used in an objective comparison of automatic segmentation algorithms for the obtained images. A statistical classification on the performance of several automatic segmentation algorithms for this tv^e of images is reached.展开更多
基金partially supported by the Plan Nacional de Investigación Científica, Desarrollo e Investigación Tecnológica (I+D+i) (Nos. AGL2011/25175 and AGL2015/69697P)by DGUI (Comunidad de Madrid) and UPM (No. QM100245066)
文摘A protocol for obtaining digital images from natural porous media with a wide range of pore sizes, intended for fractal studies of the porosity, is proposed. Soil porosity is used as para- digm of complex natural porous media in this study. The use of several imaging devices and fluores- cent compounds to enhance the contrast between the solid and the pore phase is tested. Finally a protocol is reached using a photo camera and a confocal microscope. It is the first time that confocal microscopy is used for this purpose. Artificial porous images are created through random Sierpinski carpet fractals and the statistical information of real soil images. These ground truth images are used in an objective comparison of automatic segmentation algorithms for the obtained images. A statistical classification on the performance of several automatic segmentation algorithms for this tv^e of images is reached.
