The applications of multispectral and morphometric image analysis to soil thin section descriptions is examined. It is shown that unsupervised classification and contrast stretching can be used to enhance and label features of interest. Morphometric measurements, allow the shape and abundance of features in thin sections to be compared and statistical relationships established. This method of analysis offers a precision beyond that which is possible using a qualitative approach. Using soil micromorphology as the main analytical technique a case study was carried out to evaluate the applicability of image analysis to an investigation of an archaeological site at Lairg in northern Scotland. The interactions of anthropogenic activity and pedogenesis since the Bronze Age is examined. The condition of the soils prior to the Bronze Age is not known because no buried soils predated this period. Evidence suggests that in freely draining situations complete podzols might have formed by this time. During the Bronze and Iron Age intensive cultivation of soils occurred with associated erosion. In areas of the site, where human activity is dated to the Post Medieval period, deepening of A horizons is apparent and the soils are maintained as Brown Podzols. In many areas where human activity stops stagnopodzols are the dominant soil type. The potential of image analysis to aid soil micromorphological descriptions is demonstrated. Contrast stretching aided a qualitative subdivision of thin section slides during the case study. Morphometric analysis confirmed a relationship between shape of voids and c/f ratios in an Iron Age buried A horizon, A PM buried A horizon and an undated deep topsoil. It is concluded that the full benefits of image analysis, when used as a routine tool to aid thin section descriptions, will only be realised when procedures become more interactive and processes can be speeded up.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:513841 |
| Date | January 1993 |
| Creators | Acott, Timothy G. |
| Publisher | University of Stirling |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://hdl.handle.net/1893/1784 |
Page generated in 0.0019 seconds