We have developed a highly accurate support vector machine (SVM) based detector capable of identifying jarosite (K, Na, H3O)Fe 3 (SO4)2(OH)6) in the visible/NIR (350-2500 nm) spectra of both laboratory specimens and rocks in Mars analogue field environments. To keep the computational complexity of the detector to a minimum, we restricted our design to an SVM with a linear kernel and a small number of support vectors. We used our generative model to create linear mixtures of end-member library spectra to train the SVM. We validated the detector on museum quality laboratory samples (97% accuracy) and field rock samples measured in both the laboratory and the field (both 88% accuracy). In the interest of technology infusion, the detector has been integrated into the CLARAty autonomous mobile robotics software architecture