Initial commit

Flexbrdf/hytools/glint/hedley_2005.py

@@ -0,0 +1,140 @@
+# -*- coding: utf-8 -*-
+"""
+HyTools:  Hyperspectral image processing library
+Copyright (C) 2021 University of Wisconsin
+
+Authors: Evan Greenberg.
+
+This program is free software: you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation, version 3 of the License.
+
+This program is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program.  If not, see <https://www.gnu.org/licenses/>.
+"""
+import numpy as np
+from scipy import stats
+from ..masks import mask_create
+
+
+def apply_hedley_2005_correction(hy_obj, data, dimension, index):
+    """
+    Glint correction algorithm following:
+
+    Hedley, J. D., Harborne, A. R., & Mumby, P. J. (2005).
+    Simple and robust removal of sun glint for mapping shallow‐water benthos.
+    International Journal of Remote Sensing, 26(10), 2107-2112.
+    """
+    # Raise exception is there is no deep water sample provided
+    if isinstance(hy_obj.glint.get('deep_water_sample'), type(None)):
+        raise KeyError("No Deep Water Sample Provided")
+
+    if 'apply_glint' not in hy_obj.mask:
+        hy_obj.gen_mask(mask_create,'apply_glint',hy_obj.glint['apply_mask'])
+
+    if hy_obj.mask['apply_glint'].sum() == 0:
+        return data
+
+    hy_obj.glint['correction_band'] = hy_obj.wave_to_band(
+        hy_obj.glint['correction_wave']
+    )
+
+    if 'hedley_slopes' not in hy_obj.ancillary:
+        hy_obj.ancillary['hedley_slopes'] = optimize_slopes(hy_obj)
+
+    if 'hedley_nir_swir_diff' not in hy_obj.ancillary:
+        hy_obj.ancillary['hedley_nir_swir_diff'] = nir_swir_diff(hy_obj)
+
+    if dimension == 'line':
+        correction = (
+            hy_obj.ancillary['hedley_nir_swir_diff'][index, :].reshape(-1, 1)
+            * hy_obj.ancillary['hedley_slopes']
+        )
+        correction[~hy_obj.mask['apply_glint'][index, :], :] = 0
+
+    elif dimension == 'column':
+        correction = (
+            hy_obj.ancillary['hedley_nir_swir_diff'][:, index].reshape(-1, 1)
+            * hy_obj.ancillary['hedley_slopes']
+        )
+        correction[~hy_obj.mask['apply_glint'][:, index], :] = 0
+
+    elif (dimension == 'band'):
+        correction = (
+            hy_obj.ancillary['hedley_nir_swir_diff']
+            * hy_obj.ancillary['hedley_slopes'][0, index]
+        )
+        correction[~hy_obj.mask['apply_glint']] = 0
+
+    elif dimension == 'chunk':
+        x1, x2, y1, y2 = index
+        corr_diff = hy_obj.ancillary['hedley_nir_swir_diff'][y1:y2, x1:x2]
+        bandnums = data.shape[2]
+        corr_diff = np.repeat(
+            corr_diff[:, :, np.newaxis],
+            bandnums,
+            axis=2
+        )
+
+        correction = corr_diff * hy_obj.ancillary['hedley_slopes']
+        correction[~hy_obj.mask['apply_glint'][y1:y2, x1:x2], :] = 0
+
+    elif dimension == 'pixels':
+        y, x = index
+
+        correction = (
+            hy_obj.ancillary['hedley_nir_swir_diff'][y, x].reshape(-1, 1)
+            * hy_obj.ancillary['hedley_slopes']
+        )
+        correction[~hy_obj.mask['apply_glint'][y, x], :] = 0
+
+    return data - correction
+
+
+def optimize_slopes(hy_obj):
+    deep_water = hy_obj.get_chunk(
+        *hy_obj.glint['deep_water_sample'][hy_obj.file_name]
+    )
+
+    deep_correction = (
+        deep_water[:, :, hy_obj.glint['correction_band']].flatten()
+    )
+
+    # Iterate through each band to find the band-slope
+    slopes = np.empty([1, len(hy_obj.wavelengths)])
+    for i, band in enumerate(hy_obj.wavelengths):
+        # Get flattened deep water sample of band
+        wave_num = np.argmin(
+            np.abs(hy_obj.wavelengths - band)
+        )
+        wave = deep_water[:, :, wave_num].flatten()
+
+        # Regress
+        (
+            slope,
+            intercept,
+            r_value,
+            p_value,
+            std_err
+        ) = stats.linregress(
+            deep_correction,
+            wave
+        )
+        slopes[0, i] = slope
+
+    return slopes
+
+
+def nir_swir_diff(hy_obj):
+    nir_swir_array = np.copy(
+        hy_obj.get_wave(hy_obj.glint['correction_wave'])
+    )
+    nir_swir_array[~hy_obj.mask['apply_glint']] = 0
+    nir_swir_min = np.percentile(nir_swir_array[nir_swir_array > 0], .0001)
+
+    return nir_swir_array - nir_swir_min