1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
|
import numpy as np
import xarray as xr
from polsarpro.util import vec_to_mat
from polsarpro.util import S_to_C3, S_to_C3_dask, S_to_C3
from polsarpro.util import S_to_T3, S_to_T3_dask, S_to_T3
from polsarpro.util import T3_to_C3, T3_to_C3_dask, T3_to_C3
from polsarpro.util import C3_to_T3, C3_to_T3_dask, C3_to_T3
from polsarpro.util import boxcar, boxcar_dask, boxcar_xarray
def test_vec_to_mat():
N = 128
D = 3
v = np.random.rand(N, N, D) + 1j * np.random.rand(N, N, D)
M = vec_to_mat(v)
assert M.shape == (N, N, D, D)
# M has to be Hermitian
assert np.allclose(M.transpose((0, 1, 3, 2)), M.conj())
assert np.allclose(M.diagonal(axis1=2, axis2=3).imag, 0)
def test_S_to_C3():
N = 128
S = np.random.rand(N, N, 2, 2) + 1j * np.random.rand(N, N, 2, 2)
for fun in [S_to_C3, S_to_C3_dask]:
C3 = fun(S)
assert C3.shape == (N, N, 3, 3)
# C3 has to be Hermitian
assert np.allclose(C3.transpose((0, 1, 3, 2)), C3.conj())
assert np.allclose(C3.diagonal(axis1=2, axis2=3).imag, 0)
# Xarray version
dims = ("y", "x")
S_dict = dict(
hh=xr.DataArray(S[..., 0, 0], dims=dims),
hv=xr.DataArray(S[..., 0, 1], dims=dims),
vh=xr.DataArray(S[..., 1, 0], dims=dims),
vv=xr.DataArray(S[..., 1, 1], dims=dims),
)
Sx = xr.Dataset(S_dict, attrs=dict(poltype="S"))
C3x = S_to_C3(Sx)
# test ouput shapes and types
assert C3x.poltype == "C3"
assert all(C3x[var].shape == (N, N) for var in C3x.data_vars)
assert all(C3x[var].dtype == "float32" for var in ["m11", "m22", "m33"])
assert all(C3x[var].dtype == "complex64" for var in ["m12", "m13", "m23"])
def test_S_to_T3():
N = 128
S = np.random.rand(N, N, 2, 2) + 1j * np.random.rand(N, N, 2, 2)
for fun in [S_to_T3, S_to_T3_dask]:
T3 = fun(S)
assert T3.shape == (N, N, 3, 3)
# C3 has to be Hermitian
assert np.allclose(T3.transpose((0, 1, 3, 2)), T3.conj())
assert np.allclose(T3.diagonal(axis1=2, axis2=3).imag, 0)
# Xarray version
dims = ("y", "x")
S_dict = dict(
hh=xr.DataArray(S[..., 0, 0], dims=dims),
hv=xr.DataArray(S[..., 0, 1], dims=dims),
vh=xr.DataArray(S[..., 1, 0], dims=dims),
vv=xr.DataArray(S[..., 1, 1], dims=dims),
)
Sx = xr.Dataset(S_dict, attrs=dict(poltype="S"))
T3x = S_to_T3(Sx)
# test ouput shapes and types
assert T3x.poltype == "T3"
assert all(T3x[var].shape == (N, N) for var in T3x.data_vars)
assert all(T3x[var].dtype == "float32" for var in ["m11", "m22", "m33"])
assert all(T3x[var].dtype == "complex64" for var in ["m12", "m13", "m23"])
def test_T3_to_C3():
N = 128
D = 3
v = np.random.rand(N, N, D) + 1j * np.random.rand(N, N, D)
# fake T3 matrix
T3 = vec_to_mat(v)
for fun in [T3_to_C3, T3_to_C3_dask]:
C3 = fun(T3)
assert C3.shape == (N, N, 3, 3)
# C3 has to be Hermitian
assert np.allclose(C3.transpose((0, 1, 3, 2)), C3.conj())
assert np.allclose(C3.diagonal(axis1=2, axis2=3).imag, 0)
# span must remain the same
assert np.allclose(
C3.diagonal(axis1=2, axis2=3).sum().real,
T3.diagonal(axis1=2, axis2=3).sum().real,
)
# Xarray version
dims = ("y", "x")
T3_dict = dict(
m11=xr.DataArray(T3[..., 0, 0].real.astype("float32"), dims=dims),
m22=xr.DataArray(T3[..., 1, 1].real.astype("float32"), dims=dims),
m33=xr.DataArray(T3[..., 2, 2].real.astype("float32"), dims=dims),
m12=xr.DataArray(T3[..., 0, 1].astype("complex64"), dims=dims),
m13=xr.DataArray(T3[..., 0, 2].astype("complex64"), dims=dims),
m23=xr.DataArray(T3[..., 1, 2].astype("complex64"), dims=dims),
)
T3x = xr.Dataset(T3_dict, attrs=dict(poltype="T3"))
C3x = T3_to_C3(T3x)
# test ouput shapes and types
assert C3x.poltype == "C3"
assert all(C3x[var].shape == (N, N) for var in C3x.data_vars)
assert all(C3x[var].dtype == "float32" for var in ["m11", "m22", "m33"])
assert all(C3x[var].dtype == "complex64" for var in ["m12", "m13", "m23"])
def test_C3_to_T3():
N = 128
D = 3
v = np.random.rand(N, N, D) + 1j * np.random.rand(N, N, D)
# fake T3 matrix
C3 = vec_to_mat(v)
for fun in [C3_to_T3, C3_to_T3_dask]:
T3 = fun(C3)
assert T3.shape == (N, N, 3, 3)
# T3 has to be Hermitian
assert np.allclose(T3.transpose((0, 1, 3, 2)), T3.conj())
assert np.allclose(T3.diagonal(axis1=2, axis2=3).imag, 0)
# span must remain the same
assert np.allclose(
T3.diagonal(axis1=2, axis2=3).sum().real,
C3.diagonal(axis1=2, axis2=3).sum().real,
)
# Xarray version
dims = ("y", "x")
C3_dict = dict(
m11=xr.DataArray(C3[..., 0, 0].real.astype("float32"), dims=dims),
m22=xr.DataArray(C3[..., 1, 1].real.astype("float32"), dims=dims),
m33=xr.DataArray(C3[..., 2, 2].real.astype("float32"), dims=dims),
m12=xr.DataArray(C3[..., 0, 1].astype("complex64"), dims=dims),
m13=xr.DataArray(C3[..., 0, 2].astype("complex64"), dims=dims),
m23=xr.DataArray(C3[..., 1, 2].astype("complex64"), dims=dims),
)
C3x = xr.Dataset(C3_dict, attrs=dict(poltype="C3"))
T3x = C3_to_T3(C3x)
# test ouput shapes and types
assert T3x.poltype == "T3"
assert all(T3x[var].shape == (N, N) for var in T3x.data_vars)
assert all(T3x[var].dtype == "float32" for var in ["m11", "m22", "m33"])
assert all(T3x[var].dtype == "complex64" for var in ["m12", "m13", "m23"])
def test_boxcar():
N = 128
D = 3
v = np.random.rand(N, N, D) + 1j * np.random.rand(N, N, D)
M = vec_to_mat(v)
for fun in [boxcar, boxcar_dask]:
param = {"img": M, "dim_az": 5, "dim_rg": 3}
M_box = fun(**param)
assert M_box.shape == M.shape
assert M_box.dtype == M.dtype
# output has to be Hermitian
assert np.allclose(M_box.transpose((0, 1, 3, 2)), M_box.conj())
assert np.allclose(M_box.diagonal(axis1=2, axis2=3).imag, 0)
# Xarray version
dims = ("y", "x")
M_dict = dict(
m11=xr.DataArray(M[..., 0, 0].real.astype("float32"), dims=dims),
m22=xr.DataArray(M[..., 1, 1].real.astype("float32"), dims=dims),
m33=xr.DataArray(M[..., 2, 2].real.astype("float32"), dims=dims),
m12=xr.DataArray(M[..., 0, 1].astype("complex64"), dims=dims),
m13=xr.DataArray(M[..., 0, 2].astype("complex64"), dims=dims),
m23=xr.DataArray(M[..., 1, 2].astype("complex64"), dims=dims),
)
Mx = xr.Dataset(M_dict)
param = {"img": Mx, "dim_az": 5, "dim_rg": 3}
M_box = boxcar_xarray(**param)
# test ouput shapes and types
assert all(Mx[var].shape == (N, N) for var in Mx.data_vars)
assert all(Mx[var].dtype == "float32" for var in ["m11", "m22", "m33"])
assert all(Mx[var].dtype == "complex64" for var in ["m12", "m13", "m23"])
|
