Newer
Older
Paolo Veglio
committed
import numpy as np
from numpy.lib.stride_tricks import sliding_window_view
_bad_data = -999.0
def spatial_var(rad, threshold):
test = sliding_window_view(np.pad(rad, [1, 1], mode='constant'), (3, 3)) - np.expand_dims(rad, (2, 3))
test = np.abs(test.reshape(rad.shape[0], rad.shape[1], 9))
var = np.ones(test.shape)
var[test < threshold] = 0
return var.sum(axis=2)
def sunglint(viirs_data, threshold, bit, conf):
m09 = viirs_data.M02.values
m20 = viirs_data.M12.values
m31 = viirs_data.M15.values
m02 = viirs_data.M07.values
Paolo Veglio
committed
m01 = viirs_data.M05.values
Paolo Veglio
committed
irclr = np.ones(viirs_data.M02.shape)
Paolo Veglio
committed
idx = np.nonzero((m02 > 1.3) & (m01 <= 2.0))
m02[idx] = 1.3
idx = np.nonzero((m02 < -99) | (m02 > 1.3))
m02[idx] = _bad_data
irclr[bit < 3] = 0
Paolo Veglio
committed
var = spatial_var(m31, 0.4)
reg_var_mean = sliding_window_view(np.pad(m02, [1, 1], mode='constant'),
(3, 3)).reshape(m02.shape[0], m02.shape[1], 9).mean(axis=2)
reg_std = sliding_window_view(np.pad(m02, [1, 1], mode='constant'),
(3, 3)).reshape(m02.shape[0], m02.shape[1], 9).std(axis=2)
d37_11 = m20 - m31
idx = np.nonzero((var == 0) & (m09 != _bad_data) & (m20 != _bad_data) & (m31 != _bad_data) & (irclr == 1) &
(m02 != _bad_data) & (m09 < threshold['sngm09']) & (d37_11 >= threshold['sg_tbdfl']))
conf[idx] = 0.67
idx = np.nonzero((var == 0) & (m09 != _bad_data) & (m20 != _bad_data) & (m31 != _bad_data) & (irclr == 1) &
(m02 != _bad_data) & (m09 < threshold['sngm09']) & (d37_11 >= threshold['sg_tbdfl']) &
(reg_var_mean != _bad_data) & (reg_var_mean*reg_std < threshold['sngm02vm']))
conf[idx] = 0.96
return conf
def spatial(viirs_data, threshold, scene, conf):
Paolo Veglio
committed
m02 = viirs_data.M07.values
Paolo Veglio
committed
m31 = viirs_data.M15.values
var_m31 = spatial_var(m31, 0.40)
var_m02 = spatial_var(m02, 0.0020)
idx = np.nonzero((conf > 0.95) & (scene['day'] == 1) & (var_m31 == 0) & (var_m02 == 0))
conf[idx] = 1
idx = np.nonzero((conf > 0.66) & (conf <= 0.95) & (scene['day'] == 1) & (var_m31 == 0) & (var_m02 == 0))
conf[idx] = 0.96
idx = np.nonzero((conf <= 0.66) & (scene['day'] == 1) & (var_m31 == 0))
conf[idx] = 0.67
return conf
Paolo Veglio
committed
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
def land(viirs_data, threshold, scene, conf):
m04 = viirs_data.M04.values.ravel()
m05 = viirs_data.M08.values.ravel()
m20 = viirs_data.M12.values.ravel()
m22 = viirs_data.M13.values.ravel()
m31 = viirs_data.M15.values.ravel()
eco = viirs_data.eco.values.ravel()
desert = scene['desert'].ravel()
conf = conf.ravel()
tbadj = 0
ldsbt11bd = np.array(threshold['Land_Restoral']['ldsbt11bd'])
ldsbt11 = np.array(threshold['Land_Restoral']['ldsbt11bd'])
irclr = 1
hds11 = np.ones((eco.shape[0], 3)) * (ldsbt11 - tbadj)
hds11[eco == 8, :] = ldsbt11bd - tbadj
if irclr == 1:
conf[m31 > hds11[:, 2]] = 1
conf[(m31 > hds11[:, 1]) & (m31 <= hds11[:, 2])] = 0.96
conf[m31 <= hds11[:, 1]] = 0.5
m5_4_thr = np.full(eco.shape, threshold['Land_Restoral']['ldr5_4_thr'])
m5_4_thr[desert == 1] = threshold['Land_Restoral']['ldsr5_4_thr']
m5_4 = m05/m04
md1 = m20 - m22
md2 = m22 - m31
idx = np.nonzero((md1 < threshold['Land_Restoral']['ld20m22']) &
(md2 < threshold['Land_Restoral']['ld22m31']) &
(m5_4 > m5_4_thr) &
(conf <= 0.95))
conf[idx] = 0.96
conf = conf.reshape(viirs_data.M01.shape)
return conf
def coast(viirs_data, threshold, scene, conf):
m01 = viirs_data.M05.values
m02 = viirs_data.M07.values
coast_ndvi = threshold['Coastal_NDVI_Thresholds']['coast_ndvi']
irclr = 1
if irclr == 1:
ndvi = (m02 - m01)/(m01 + m02)
idx = np.nonzero((ndvi <= coast_ndvi[0]) | (ndvi >= coast_ndvi[1]))
conf[idx] = 1
return conf