Estimation of sedimentary rock grain size distributions in drill core is a task routinely performed by geologists to log changes to the size and sorting of the sedimentary rock grains. This information is valuable for interpreting the primary depositional processes responsible for the emplacement of sediment, their likely geometry and resultant (measureable) flow properties. Currently, this is performed by physical examination of the drill core, whereby the grain size of a section of core is estimated (through visual comparison with a card illustrating reference grain sizes) and the depths of bed boundaries of changes in grain size are logged by a geologist.
Examination of a slabbed drill core involves the visual inspection of a surface (or 2D slice) that passes through sedimentary grains at various positions and angles. Comparison of the long-axis length distributions of grains to standardised grain size charts provides a measure of grain size and size distribution.
Although robust grain size measurements can be made via destructive techniques (i.e. crushing the sedimentary rock followed by laser sieve analysis), most publically available drill core must not be damaged and ideally should not be picked up. As such, visual techniques of grain size estimation must be employed.
A previous method for automatically estimating the grain size distribution of a section of core has used special sample preparation and imaging conditions, e.g. viewing thin sections of the core under cross-polarised light. Handling and destruction of the drill core to make thin sections is a time consuming process, and destruction of the drill core is not always allowed.
The aim of this project is to conduct a feasibility study into the use of a portable image capture and processing method for the estimation of grain size distributions from slabbed core. A high resolution digital camera with macro lens will be used to image the core non-destructively within the core trays. We will investigate various image segmentation and texture analysis techniques to characterise grain size distributions from these images. This research will be performed with the goals of producing a solution for fast, non-destructive and repeatable grain size analysis, and with an aim to be incorporated in a later project into a portable system for automated estimation of the grain size distribution that can be used to examine cores stored at the GSWA core library.