Scene illumination in digital photography is of limited types, including Daylights of various temperatures, Horizon, Incandescent and various Fluorescents. For white balancing purposes in consumer digital photography, it is often sufficient to compute the scene illumination with limited accuracy. Therefore, rather than estimating the tristimulus color value or the illumination spectrum as a continuous (vector) variable, we can detect the illumination via hypothesis testing and selecting from a finite set of candidate illuminations. An illumination detection approach is introduced in this work by incorporating linear decomposition of the spectral surface reflectance.Illumination estimation, illumination detection, hypothesis testing, linear decomposition, white balancing.