The main aim of this research was to investigate the possible uses of digital image scanning to assess the colour and colour fastness of textiles. This method has the potential to be a simpler alternative to camera based colour fastness assessment, for example'by the DigiEye system.. The scanning device used was an Epson Perfection 2400 Photo scanner, a mid-range quality device. The scanner was operated in· two different modes, 24 and 48 bits colour resolution and was calibrated by adjusting its setting, testing its repeatability for colour measurement and examining the spatial uniformity of response of the scanner. The results showed that the spatial uniformity of the 'scanner response was excellent for the majority of the scanning area but less uniform for the areas along its scanning borders. The repeatability of the scanner values for the colour measurement purposes was reasonable. The colour and colour fastness determination by the 24 and the 48 bits scanner modes indicated that there were no practical benefits to the use of the 48 bits. The colorimetric response of the scanner was characterised with the XYZ and the L*a*b* approaches in which the RGB responses of the scanner were transformed to the XYZ tristimulus values and to the L*a*b* values, respectively. It was shown that pre-treatment of the data by linearisation against either luminance or mean reflectance gave very good results with the XYZ approach to scanner. characterisation and that linearisation against L* gave excellent results for the L*a*b* approach to scanner characterisation. In the characterisation of the scanner two different techniques for parameter optimisation were applied, regression and iteration. The experiments show that the iteration techniques marginally improved the performance of the XYZ approach but it had no significant effect on the L*a*b* approach. The experiments conducted to find the best colour chart to use for colorimetric characterisation of the scanner showed that the TC chart which was made from an atlas of textile colour patches outperformed the CCDCchart which was made of 165 patches. of the ColorChecker DC. From the performance results of the colorimetric characterisation with different polynomials in the regression techniques it was found that using a second order polynomial to relate RGB to output colour parameters had the best generalisation and a good memorisation effect. It was concluded that for colour fastness assessment the scanner should be characterised with a regression technique fitting the coefficients of a second order polynomial us'ing the RGB to L*a*b* approach 'of colorimetric characterisation and theL* method for linearisation of the eeDe chart patches' RGB. The relative spectral power distribution (SPD) of the scanner lamp was also measured and it was concluded that part ofresidual inaccuracies of the scanner colour measurement could have arisen from the spiky shape ofthe SPD. The characterised scanner was used for instrumental colour fastness assessment of two sets of panels, one set of 60 staining and 15 colour change panels called the old set and a set of240 staining and 40 colour change panels made during the research to increase the range and uniformity ofthe colour distribution ofthe panels. An existing set of observer results were available for the old set and were obtained from 38 observers from 12 laboratories ofthe UK. The grades for the new set were obtained from 6 observers from two laboratories in the UK. The instrumental colour fastness grades of the sets were determined by the ISO, the GRS and the GRC formulae using a spectrophotometer, DigiEye system and the scanner. On the basis of the results of these devices in comparison to the average visual grading, it was c,oncluded that the ISO staining formula as well as the GRS and the GRC predict grades with smaller disagreement to the average of the visual grades than the inter observer variation. However, the ISO colour change formula determined grades with a higher level of disagreement to the average visual grade that was higher than the inter-observer variation. The scanner grading showed very good agreement to the· spectrophotometric and DigiEye grading with the ISO staining, the GRS and the GRC formulae but only moderate agreement with the ISO colour change formulae. The scanner repeatability for colour fastness assessment was determined and shown to be very good, however, as the frequency of assessment by the scanner in a research situation is not comparable to its industrial application, it was suggested to characterise the scanner once a day. The characterisation of the scanner with the research TC chart was shown to produce np significant improvement in the results that justify the difficulty of the preparation and maintemmce ofthis chart. A comparison was made between visual grades obtained from experienced and from inexperienced assessors of colour fastness. The results suggest that the grades given by the inexperienced are comparable to those assessed by the experienced assessors and the disagreements between the average grades are less than the experienced assessors inter-observer variation. However, as the experimental conditions for the set of the inexperienced assessors were the same but .were not the same for the member of the set of the experienced assessors it was suggested that further study is needed.