The linearization of camera signals and spectral data is a significant step in the color characterization of image capturing systems. Even though output signals from camera detector usually have a reasonable linear relationship with incident spectral radiance, several factors might lead to slight deviations from perfect linearity. Differences in capturing geometries can be a source of non-linearity between these two quantities. In this research, a surface correction equation is introduced to compensate for differences between the camera signals and reflectance measurement with the aim of improving linearity. We utilized the Saunderson surface correction to account for boundary reflections based on measurement geometries. To investigate the idea, three experimental phases were set up. In the first experiment, the reflectance data from two different spectrophotometers were compared with those from a spectroradiometer in two dissimilar lighting conditions. According to the results, the Saunderson equation is capable of improving the measured reflectances from spectrophotometers to be fit to the actual spectral radiances from the spectroradiometer independent of capturing and lighting geometry. In the second phase of the experiment, a digital still camera was characterized using the measured and surface-corrected spectral reflectances. Finally, in the third phase of the experiment, a ColorChecker SG was imaged and used as independent verification data. According to the results, the surface correction improved the linear correlation of spectral reflectances and camera signals for all geometries and spectral data.