IMIS | Flanders Marine Institute

Flanders Marine Institute

Platform for marine research


Publications | Institutes | Persons | Datasets | Projects | Maps
[ report an error in this record ]basket (0): add | show Printer-friendly version

Prediction of sediment properties in core samples using image analysis of radiographs
Dung, H.T. (1999). Prediction of sediment properties in core samples using image analysis of radiographs. MSc Thesis. Vrije Universiteit Brussel: Brussel. 115 pp.

Thesis info:
    Vrije Universiteit Brussel; Faculteit Wetenschappen; Vakgroep Biologie; Ecological Marine Management Programme (ECOMAMA), more

Available in  Author 
  • VLIZ: Archive A.THES13 [17432]
  • VLIZ: Non-open access 228374
Document type: Dissertation


Author  Top 
  • Dung, H.T.

    This study presents a simple, relatively inexpensive and fast image analysis technique on radiographs of core samples for sedimentological study. The technique was developed from the idea that there are relationships between the results of non-destructive and destructive analyses on core samples. Non-destructive analyses include the bulk density measurement and X-ray radiography. Destructive analyses were carried out including grain-size and organic matter content determination. Two sediment cores numbered 97B35 and 98B69 taken from the Tielrode marsh in the Schelde estuary in 1997 and 1998 respectively were used for the purpose of this study. The relationships between grayscale values on the radiograph and bulk density, grain-size parameters, and organic matter content were examined and built up based on the non-destructively and destructively analysed results of the core 97B35. Statistical analysis shows that there are strong correlations between grayscale values on the radiograph and bulk density (correlation coefficient: r = 0.96), between grayscale values and same grain-size parameters such as sand (r = 0.91), clay (r = -0.90) fractions, and median grain (r = 0.89) in the core 97B35. A weaker correlation has been found between grayscale value and orgarnic matter content (r = -0.75). The strong correlations between grayscale value and bulk density, grayscale value and some grain-size parameters indicate that bulk density and grain-size parameters can be roughly predicted from the grayscale values obtained by the quantitative image analysis on the corresponding image areas of the radiograph. Prediction was done on the core 98B69 as a demonstration of the study technique. For the improvement of this technique, two methods are suggested including reducing sample thickness and decreasing the maximum angle of incidence of X-rays to core stratification.

All data in IMIS is subject to the VLIZ privacy policy Top | Author