The NHI (Normalized Hotspot Indices; Marchese et al., 2019) algorithm uses two normalized indices to identify thermal anomalies through Landsat-8 OLI and Sentinel-2 MSI data:
NHISWIR= (L2.2-L1.6)/(L2.2+L1.6)
NHISWNIR= (L1.6-L0.8)/(L1.6+L0.8)
where, L2.2, L1.6, and L0.8 are the TOA radiaces [W⸱m-2⸱sr-1⸱m-1] measured, for each pixel of the scene, at around 2.2 µm (SWIR2), 1.6 µm (SWIR1), and 0.8 µm (NIR). An initial test on SWIR2 radiance and a spectral filter, for multi-spectral misregistration issues of Sentinel-2 MSI imagery, are used before the computation of the two normalized indices to minimize false positives. Pixels having values of NHISWIR > 0 OR NHISWNIR > 0 are flagged as "hot" by the algorithm.