import numpy as np from pylst.exceptions import assert_required_keywords_provided from pylst.utils import fractional_vegetation_cover # Credit to the original author: Oladimeji Mudele (pylandtemp) # Original source: https://github.com/pylandtemp/pylandtemp class SplitWindowParentLST: def __init__(self): # Maximum Earth temperature for clipping the calculated LST self.max_earth_temp = 273.15 + 56.7 def __call__(self, **kwargs) -> np.ndarray: # Calculate LST and clip values exceeding the maximum Earth temperature lst = self._compute_lst(**kwargs) lst[lst > self.max_earth_temp] = np.nan return lst def _compute_lst(self, **kwargs): # Abstract method for computing LST, to be implemented in derived classes raise NotImplementedError("Concrete method yet to be implemented") class SplitWindowJiminezMunozLST(SplitWindowParentLST): cwv = 0.013 def _compute_lst(self, **kwargs) -> np.ndarray: # Concrete implementation for Jiminez-Munoz method required_keywords = [ "emissivity_10", "emissivity_11", "brightness_temperature_10", "brightness_temperature_11", "mask", ] assert_required_keywords_provided(required_keywords, **kwargs) # Extracting required input parameters tb_10 = kwargs["brightness_temperature_10"] tb_11 = kwargs["brightness_temperature_11"] emissivity_10 = kwargs["emissivity_10"] emissivity_11 = kwargs["emissivity_11"] mask = kwargs["mask"] # Computing LST using Jiminez-Munoz method mean_e = (emissivity_10 + emissivity_11) / 2 diff_e = emissivity_10 - emissivity_11 diff_tb = tb_10 - tb_11 lst = ( tb_10 + (1.387 * diff_tb) + (0.183 * (diff_tb ** 2)) - 0.268 + ((54.3 - (2.238 * self.cwv)) * (1 - mean_e)) + ((-129.2 + (16.4 * self.cwv)) * diff_e) ) lst[mask] = np.nan return lst class SplitWindowKerrLST(SplitWindowParentLST): """ Method reference: Kerr Y, Lagouarde J, Nerry F, Ottlé C (2004) Land surface temperature retrieval techniques and applications: case of the AVHRR. Thermal remote sensing in land surface processing. CRC Press, New York, pp 55–131 """ def _compute_lst(self, **kwargs) -> np.ndarray: """ Computes the Land Surface Temperature (LST) using the Split Window method based on Kerr et al. (2004). Args: brightness_temperature_10 (np.ndarray): Brightness temperature image obtained for band 10 brightness_temperature_11 (np.ndarray): Brightness temperature image obtained for band 11 ndvi (np.ndarray): NDVI image mask (np.ndarray[bool]): Mask image. Output will have NaN value where the mask is True. Returns: np.ndarray: Land surface temperature image """ # Ensure that required keyword arguments are provided required_keywords = [ "brightness_temperature_10", "brightness_temperature_11", "ndvi", "mask", ] assert_required_keywords_provided(required_keywords, **kwargs) # Retrieve required input data tb_10 = kwargs["brightness_temperature_10"] tb_11 = kwargs["brightness_temperature_11"] ndvi = kwargs["ndvi"] mask = kwargs["mask"] # Calculate fractional vegetation cover (pv) using NDVI pv = fractional_vegetation_cover(ndvi) # Compute Land Surface Temperature (LST) using the Kerr et al. (2004) formula lst = ( (tb_10 * ((0.5 * pv) + 3.1)) + (tb_11 * ((-0.5 * pv) - 2.1)) - ((5.5 * pv) + 3.1) ) # Mask out invalid values based on the provided mask lst[mask] = np.nan return lst class SplitWindowMcMillinLST(SplitWindowParentLST): """ Method reference: McMillin LM (1975) Estimation of sea surface temperatures from two infrared window measurements with different absorption. J Geophys Res 80(36):5113–5117. https://doi.org/10.1029/JC080i036p05113 """ def _compute_lst(self, **kwargs) -> np.ndarray: """ Computes the land surface temperature using McMillin's method. Args: **brightness_temperature_10 (np.ndarray): Brightness temperature image obtained for band 10 **brightness_temperature_11 (np.ndarray): Brightness temperature image obtained for band 11 **mask (np.ndarray[bool]): Mask image. Output will have NaN values where the mask is True. Returns: np.ndarray: Land surface temperature image """ # Check if required keywords are provided required_keywords = [ "brightness_temperature_10", "brightness_temperature_11", "mask", ] assert_required_keywords_provided(required_keywords, **kwargs) # Extracting input variables tb_10 = kwargs["brightness_temperature_10"] tb_11 = kwargs["brightness_temperature_11"] mask = kwargs["mask"] # McMillin's land surface temperature computation lst = (1.035 * tb_10) + (3.046 * (tb_10 - tb_11)) - 10.93 # Masking areas where the mask is True lst[mask] = np.nan return lst class SplitWindowPriceLST(SplitWindowParentLST): """ Method reference: Price JC (1984) Land surface temperature measurements from the split window channels of the NOAA advanced very high-resolution radiometer. J Geophys Res 89:7231–7237. https://doi.org/10.1029/JD089iD05p07231 """ def _compute_lst(self, **kwargs) -> np.ndarray: """ Computes Land Surface Temperature (LST) using the split window Price method. Args: **kwargs (dict): Keyword arguments containing necessary input arrays. Required keywords: "emissivity_10", "emissivity_11", "brightness_temperature_10", "brightness_temperature_11", "mask". Returns: np.ndarray: Land Surface Temperature (LST) array. """ required_keywords = [ "emissivity_10", "emissivity_11", "brightness_temperature_10", "brightness_temperature_11", "mask", ] assert_required_keywords_provided(required_keywords, **kwargs) # Extracting required arrays from kwargs tb_10 = kwargs["brightness_temperature_10"] tb_11 = kwargs["brightness_temperature_11"] emm_10 = kwargs["emissivity_10"] emm_11 = kwargs["emissivity_11"] mask = kwargs["mask"] # Computing LST using the Price method formula lst = (tb_10 + 3.33 * (tb_10 - tb_11)) * ((5.5 - emm_10) / 4.5) + ( 0.75 * tb_11 * (emm_10 - emm_11) ) # Masking invalid values in the result array lst[mask] = np.nan return lst class SplitWindowSobrino1993LST(SplitWindowParentLST): """ Method reference: Sobrino JA, Caselles V, Coll C (1993) Theoretical split window algorithms for determining the actual surface temperature. I1Nuovo Cimento 16:219–236. https://doi.org/10.1007/BF02524225 This class implements the Split Window algorithm based on the Sobrino 1993 method. The algorithm computes the land surface temperature (LST) using brightness temperatures and emissivity values for Landsat 8 bands 10 and 11. """ def _compute_lst(self, **kwargs) -> np.ndarray: # Define the required keywords for the algorithm required_keywords = [ "emissivity_10", "emissivity_11", "brightness_temperature_10", "brightness_temperature_11", "mask", ] # Ensure that all required keywords are provided assert_required_keywords_provided(required_keywords, **kwargs) # Retrieve necessary data from the provided keyword arguments tb_10 = kwargs["brightness_temperature_10"] tb_11 = kwargs["brightness_temperature_11"] emm_10 = kwargs["emissivity_10"] emm_11 = kwargs["emissivity_11"] mask = kwargs["mask"] # Calculate the temperature difference between tb_10 and tb_11 diff_tb = tb_10 - tb_11 # Calculate the emissivity difference between emm_10 and emm_11 diff_e = emm_10 - emm_11 # Compute the land surface temperature (LST) using the Sobrino 1993 formula lst = ( tb_10 + (1.06 * diff_tb) + (0.46 * diff_tb ** 2) + (53 * (1 - emm_10)) - (53 * diff_e) ) # Mask invalid pixels based on the provided mask, setting their values to NaN lst[mask] = np.nan return lst