This study unravels the spectral footprint of the Paraná deposit, northeastern Brazil, and provides strategies for emerald exploration in mafic- and/or ultramafic-hosted deposits using point spectral data and hyperspectral imaging. Potential host rocks (phlogopite and actinolite-phlogopite schists) were discriminated from other associated lithotypes after petrography and whole-rock geochemistry, using binary MgO versus Al2O3 and PC1 versus PC3 diagrams. Spectrally, phlogopite schist is marked by Fe-OH absorption at ~2,250 nm and Mg-OH absorptions at ~2,330 and ~2,388 nm. When actinolite is present, a shoulder at ~2,296 nm is recorded, and the first Mg-OH absorption feature is shifted to ~2,315 nm. Emerald crystals have their spectral signature attenuated in the presence of phlogopite. In a 5% emerald and 95% schist mixture simulation, two emerald absorption features are recognizable at ~1,150 (H2O absorption) and ~2,155 nm. The emerald indices EI1, EI2, and EI3 were chosen for the automated identification of emeralds in the Paraná samples based on the features at ~1,150, ~1,460, and ~2,155 nm. For the host schists, three indices were provided (MIdepth, MIratio, and ACI) considering the Mg-OH features of phlogopite and actinolite, allowing automated distinction between potential mineralized and other associated rocks of the Paraná deposit. Our results show that, operating in high-resolution mode (pixels of ~1 mm), imaging spectroscopy can detect the presence and characterize emerald in handpicked samples and drill cores based on chromium- and iron-derived absorption features in the visible to near infrared range. It thus represents a swift, reliable, nondestructive, and relatively cost-effective method for exploring for emerald and, potentially, other gems.