A technique to quantify the fabric of soils and sediments involving image processing and analysis is described in this paper. Such techniques are becoming popular in the Earth sciences, but, unlike most other studies which investigate the porosity of coarse-grained sediments, this method examines the fine porosity between clay size particles to be investigated. The usual problems associated with thresholding a grey-level image to provide the necessary binary image for analysis are partly overcome by the use of Wiener filtering methods to remove much of the degradation inherent in all imaging systems. A new algorithm based on the relative contrast histogram method is then used to determine the optimum value at which to set the threshold. The micro-porosity determined is in good agreement with gravimetric porosity measurements. Unlike normal methods, once the procedure has been initiated, no further action is required from the operator and the method can thus be used for batch processing of large numbers of images thereby enhancing the statistical reliability of the results.
Extensions of the technique allow for the variation in micro-porosity from one region of a specimen to another to be studied. Further, by segmenting the original image into domains of sub-parallel particles, it is possible to examine the intra-domain porosity. Contrary to expectations, the porosity in horizontal domains is greater than that in the vertical domains.