Traditional methods and a lack of consensus and quality control in the delimitation of lithological units lead to differences and offsets at the contact zones between map sheets. On some geological maps, such inconsistencies considerably hinder exploration work. This study evaluates the contribution of remote sensing combined with a geoscientific knowledge base to produce bedrocks maps in a geologically complex and semi-arid region of the Peruvian Andes. The region is characterized by the presence of vegetation, snow and hydrothermal alteration. The results demonstrate the potential of ASTER-TIR emissivity data to discriminate lithological units characterized by differences in silica content. These units correspond to felsic plutonic, intermediate volcanic and sedimentary carbonate rocks with a map accuracy of 72% when compared to the available geological map. Specifically with regards to quartz zones, mapping by the two methods resulted in a comparative precision of 99%. The methodological approach also demonstrated that the potential of SWIR spectral bands is sufficiently high (precision of 84%) to detect many of the minerals associated with argillic and advanced argillic alteration in the study area.