""" This code is part of the Python PolSARpro software: "A re-implementation of selected PolSARPro functions in Python, following the scientific recommendations of PolInSAR 2021" developed within an ESA funded project with SATIM. Author: Olivier D'Hondt, 2026. Scientific advisors: Armando Marino and Eric Pottier. Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0 Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License. ----- # Description: module containing physical inversion functions """ import dask.array as da import numpy as np import xarray as xr from polsarpro.auxil import validate_dataset from polsarpro.util import C4_to_C3, S_to_C3, T3_to_C3, T4_to_C3 def dubois_surface_inversion( input_data: xr.Dataset, incidence_angle: xr.DataArray, freq_ghz: float, thresh1: float, thresh2: float, calibration_coeff: float | None = None, ) -> xr.Dataset: """Run the Dubois surface inversion on a PolSAR covariance dataset. The function accepts a single-scattering input product in any of the supported polarimetric representations, converts it to C3 form when needed, and applies the Dubois empirical model using the supplied incidence-angle raster. Args: input_data (xr.Dataset): Input polarimetric dataset. Supported products are: - "S": Sinclair scattering matrix - "C3": Lexicographic covariance matrix (3x3) - "T3": Pauli coherency matrix (3x3) - "C4": 4x4 covariance matrix - "T4": 4x4 coherency matrix The dataset must share the same spatial grid as ``incidence_angle``. incidence_angle (xr.DataArray): Incidence angle raster in radians. Values must be numeric and in the range ``(0, pi/2)``. freq_ghz (float): Radar center frequency in GHz. thresh1 (float): Maximum allowed ``HV / VV`` ratio in dB for the Dubois validity mask. thresh2 (float): Maximum allowed ``HH / VV`` ratio in dB for the Dubois validity mask. calibration_coeff (float or None, optional): Optional multiplicative calibration coefficient. When provided, the HH, VV, and HV channels are scaled by ``sin(theta) / calibration_coeff`` before the inversion is applied. Defaults to None. Returns: xr.Dataset: Dataset containing the Dubois estimates and masks: ``dubois_ks`` (surface roughness parameter), ``dubois_er`` (relative dielectric constant), ``dubois_mv`` (volumetric moisture estimate), ``dubois_mask_in`` (input validity mask), ``dubois_mask_out`` (model validity mask), and ``dubois_mask_valid_in_out`` (combined mask). Notes: The returned dataset preserves the input coordinates and uses the input spatial dimensions. Output variables are stored as ``float32`` arrays. """ if not isinstance(incidence_angle, xr.DataArray): raise TypeError("incidence_angle must be an xarray.DataArray.") if not np.issubdtype(incidence_angle.dtype, np.number): raise TypeError("incidence_angle must contain numeric values.") if not isinstance(freq_ghz, (int, float, np.number)): raise TypeError(f"freq_ghz must be a number, got {type(freq_ghz).__name__}.") if freq_ghz <= 0: raise ValueError(f"freg_ghz must be strictly positive, got {freq_ghz}.") if not isinstance(thresh1, (int, float, np.number)): raise TypeError(f"thresh1 must be a number, got {type(thresh1).__name__}.") if not isinstance(thresh2, (int, float, np.number)): raise TypeError(f"thresh2 must be a number, got {type(thresh2).__name__}.") if calibration_coeff is not None: if not isinstance(calibration_coeff, (int, float, np.number)): raise TypeError( "calibration_coeff must be a number or None, " f"got {type(calibration_coeff).__name__}." ) if calibration_coeff <= 0: raise ValueError( f"calibration_coeff must be strictly positive, got {calibration_coeff}." ) allowed_poltypes = ("S", "C3", "T3", "C4", "T4") poltype = validate_dataset(input_data, allowed_poltypes=allowed_poltypes) converters = { "C3": lambda ds: ds, "T3": T3_to_C3, "C4": C4_to_C3, "T4": T4_to_C3, "S": S_to_C3, } C3 = converters[poltype](input_data) out = _apply_dubois_inversion( theta=incidence_angle, f0=freq_ghz, hh=C3.m11, vv=C3.m33, hv=C3.m22 / 2.0, calib=calibration_coeff, thresh1=thresh1, thresh2=thresh2, ) original_non_nan_mask = input_data.to_array().notnull().all("variable") out["dubois_mask_valid_in_out"] = ( out["dubois_mask_in"] * out["dubois_mask_out"] * original_non_nan_mask.astype(np.float32, copy=False) ).astype(np.float32, copy=False) return xr.Dataset( {k: (tuple(input_data.dims), v.data) for k, v in out.items()}, attrs={ "poltype": "dubois_surface_inversion", "description": "Results of the Dubois surface inversion.", }, coords=input_data.coords, ) # helper function, do not use directly def _apply_dubois_inversion(theta, f0, hh, vv, hv, calib, thresh1, thresh2): scale = np.sin(theta) / calib if calib is not None else 1.0 hh = hh * scale vv = vv * scale hv = hv * scale lambd = 100 * 0.3 / f0 eps = np.finfo(np.float32).eps theta_valid = np.isfinite(theta) & (theta > 0) & (theta < (np.pi / 2.0)) hh_pos = hh > 0 vv_pos = vv > 0 base_valid = theta_valid & hh_pos & vv_pos vv_safe = xr.where(vv_pos, vv, 1.0) hh_safe = xr.where(hh_pos, hh, 1.0) theta_safe = xr.where(theta_valid, theta, 1.0) msk_valid = ( base_valid & (hv / vv_safe < 10 ** (thresh1 / 10.0)) & (hh / vv_safe < 10 ** (thresh2 / 10.0)) ) ks_inv = np.exp( 1.36905 * np.log(hh_safe) - 0.83333 * np.log(vv_safe) + 0.446425 * np.log(np.cos(theta_safe)) + 3.34525 * np.log(np.sin(theta_safe)) - 0.375 * np.log(lambd) + 1.78989 * np.log(10) ) msk_ks = base_valid & (ks_inv >= 0) & (ks_inv <= np.pi) ks_dub = xr.where(msk_valid & msk_ks, ks_inv, 0.0) ks_safe = xr.where(ks_dub > 0, ks_dub, np.nan) er_inv = ( np.log10(hh_safe) + np.log10(vv_safe) + 5.12 + np.log10(np.cos(theta_safe)) - 2.5 * np.log10(ks_safe + eps) - 1.4 * np.log10(lambd) ) / (0.074 * np.tan(theta_safe) + eps) msk_er = base_valid & (er_inv > 0) & (er_inv <= 100) er_dub = xr.where(msk_valid & msk_ks & msk_er, er_inv, 0.0) er_safe = xr.where(er_dub > 0, er_dub, np.nan) mv_inv = ( -5.3e-2 + 2.92e-2 * er_safe - 5.5e-4 * er_safe**2 + 4.3e-6 * er_safe**3 ) * 100 msk_mv = base_valid & (mv_inv > 0) & (mv_inv <= 100) mv_dub = xr.where(msk_valid & msk_ks & msk_er & msk_mv, mv_inv, 0.0) msk_out = (msk_valid & msk_ks & msk_er & msk_mv).astype(np.float32) return { "dubois_ks": ks_dub.astype(np.float32, copy=False), "dubois_er": er_dub.astype(np.float32, copy=False), "dubois_mv": mv_dub.astype(np.float32, copy=False), "dubois_mask_out": msk_out.astype(np.float32, copy=False), "dubois_mask_in": msk_valid.astype(np.float32, copy=False), }