This paper tackles principal component analysis (PCA) in images that include wavelengths between 380-1000 nm. Our approach is focussed on taking advantage of the potencial of ultraviolet and infrarred images, in combination with the visible ones, to improve documentation process and rock art analysis. In this way, we want to improve the discrimination between pigment and support rock, and analyse the spectral behaviour of rock art paintings in the ultraviolet and infrared regions. Three images were used, one image from the ultraviolet (UV) region, one from the visible region (VIS) and another one from the near infrared region (NIR). Optical filters coupled to the camera optics were used to take the images. These filters capture specific wavelengths excluding radiation that we are not interested in registering. Finally, PCA is applied to the acquired images. The results obtained demonstrate the PCA usefulness with imagery in this field and also it is possible to extract some conclusions about the correspondent paint pigments.http://dx.doi.org/10.4995/CIGeo2017.2017.6597