X-ray computed tomography (CT) provides nondestructive three-dimensional visualization and characterization, creating images that map the variation of X-ray attenuation within objects, which relates closely to bulk density. In this work it was considered interesting to use this technique as a tool for geotechnical inspection sample. CT scans of soil samples of Castelló d'Empúries (close to Figueres, Spain) were quantitatively evaluated to determine wet bulk density. The CT scans performed over Castello soil samples were obtained as part of a larger experimental program related to the Soft Soil Project (Arroyo et al., 2012). This project had as a main objective the evaluation of sample disturbance produced by various sampling methods. A state-of-art of X-ray computed tomography applied to geosciences was perfomed. It was found that X-ray attenuation depends on the bulk density of the sample, its porosity, water content and chemical composition in a complex way. Therefore, experimental calibration was necessary to unambiguosly determine the wet bulk density of the Castello soil from the CAT values. A complementary experimental campaign using several reconstituted soils was performed for that purpose. Once the calibration was established it was applied to the CT scans of the Castello soil samples to obtain detailed profiles of wet bulk density for the tube samples and specific values for the lab test specimens. For the late wet bulk density values had been previously obtained from lab procedures. It was the possible to check the quality of the results obtained from CT scan analysis. Very good agreement was attained between both measurements. This means that the calibration and the post-process of the CT scan data were successful. Specimen quality classification results of the Soft Soil Project were reviewed, since the initial results were unreliable. The reviewing process included the sample disturbance criterion applied (Lunne et al. 1997) and all the laboratory tests results. The newly obtained CAT density was useful to identify some mistakes in the original test interpretation. After correcting then, the results were re-analyzed. A new quality classification for the specimens was obtained and it was checked for consistency against the results of the laboratory tests. This study confirms the applicability of the Lunne criteria of sample quality well away from the geotechnical setting for which they were originally proposed. The CT scan images and the sample quality assessment results indicate that, for the Castello soil, there is no advantage in using Osterberg piston-type samplers instead of simpler pushed Shelby tubes.