Colour is an important physical property in the characterization of soil type and the description of soil profiles. Quantitative data from spectrophotometers and colorimeters have been used in soil research for this purpose, but semi-quantitative M unsell colour description remains the main method of soil colour evaluation. Low-cost digital devices (cameras and scanners) could largely replace the semi-quantitative assessment of colour by M unsell charts if such devices can be calibrated colorimetrically to provide accurate and reproducible data. Robust application of such tools, however, requires standardized light sources, which precludes the use of digital cameras as viable devices for use in the field. Flatbed scanners, on the other hand, enable 2- D imaging by a contact method under consistent lighting conditions. Power can be provided to such scanners through a USB port by a laptop computer, so they can be used as viable devices in the field. In this study, we explored the feasibility of using flatbed scanners to derive colorimetrically accurate images and data from a set of 161 soil samples. The efficacy of our approach was tested with two low-cost scanners, and included analysis of two commercial colour charts, six printed colour charts and three editions of the M unsell S oil C olour chart to assess the optimum methods of colorimetric calibration. For both scanners tested, we found that accurate colour characterization could be achieved for > 95% of the soil samples studied (i.e. with colour errors barely perceptible by the human eye). These results illustrate the merit and efficacy of this rapid and low-cost approach for soil colour evaluation.