import math
from eolearn.core import FeatureType, EOTask
import numpy as np
from eolearn.core import AddFeatureTask, RemoveFeatureTask
[docs]class VegetationIndicesVHRS(EOTask):
"""Compute vegetation indices for VHRS data (only 4 bands)"""
def __init__(self, feature_name):
self.feature_name = feature_name
[docs] def calcul_ratio_vegetation_indices(self):
self.NDVI = (self.B4 - self.B3) / (self.B4 + self.B3)
self.NDWI = (self.B2 - self.B4) / (self.B2 + self.B4)
self.VARI = (self.B2 - self.B3) / (self.B2 + self.B3 - self.B1)
def execute(self, eopatch, **kwargs):
arr0 = eopatch.data[self.feature_name]
# Raw data
self.B1 = arr0[..., 0]
self.B2 = arr0[..., 1]
self.B3 = arr0[..., 2]
self.B4 = arr0[..., 3]
# VIS
self.calcul_ratio_vegetation_indices()
add_NDVI = AddFeatureTask((FeatureType.DATA, "NDVI"))
add_NDVI.execute(eopatch=eopatch, data=self.NDVI[..., np.newaxis])
add_NDWI = AddFeatureTask((FeatureType.DATA, "NDWI"))
add_NDWI.execute(eopatch=eopatch, data=self.NDWI[..., np.newaxis])
add_VARI = AddFeatureTask((FeatureType.DATA, "VARI"))
add_VARI.execute(eopatch=eopatch, data=self.VARI[..., np.newaxis])
return eopatch
[docs]class BiophysicalParameters:
def __init__(
self,
B03,
B04,
B05,
B06,
B07,
B8A,
B11,
B12,
viewZenithMean,
sunZenithAngles,
viewAzimuthMean,
sunAzimuthAngles,
):
"""
Compute biophysical parameters in Sentinel-2 data using the code available in SNAP biophysical processor.
Parameters
----------
B03-B12 : np.array
Band from S2 image (4-d format T*H*W*1)
Illumination properties:
Band from S2 image (4-d format T*H*W*1)
"""
self.B03 = B03
self.B04 = B04
self.B05 = B05
self.B06 = B06
self.B07 = B07
self.B8A = B8A
self.B11 = B11
self.B12 = B12
self.viewZenithMean = viewZenithMean
self.sunZenithAngles = sunZenithAngles
self.viewAzimuthMean = viewAzimuthMean
self.sunAzimuthAngles = sunAzimuthAngles
@staticmethod
def _normalize(unnormalized, mini, maxi):
"""Normalize inputs Neural Network"""
return 2 * (unnormalized - mini) / (maxi - mini) - 1
@staticmethod
def _denormalize(normalized, mini, maxi):
"""Denormalize output Neural Network"""
return 0.5 * (normalized + 1) * (maxi - mini) + mini
@staticmethod
def _funccos(t):
return np.vectorize(lambda t: math.cos(t))
@staticmethod
def _neuron(
cste,
w_b03,
w_b04,
w_b05,
w_b06,
w_b07,
w_b8A,
w_b11,
w_b12,
w_viewZen_norm,
w_sunZen_norm,
w_relAzim_norm,
b03_norm,
b04_norm,
b05_norm,
b06_norm,
b07_norm,
b8a_norm,
b11_norm,
b12_norm,
viewZen_norm,
sunZen_norm,
relAzim_norm,
):
def tansig(input_neuron):
"""Activation function neural network"""
funcexp = np.vectorize(lambda t: math.exp(t))
return 2 / (1 + funcexp(-2 * input_neuron)) - 1
"""Define the neuron, which is a linear combination of the metadata of sentinel2 images"""
var_sum = (
cste
+ w_b03 * b03_norm
+ w_b04 * b04_norm
+ w_b05 * b05_norm
+ w_b06 * b06_norm
)
var_sum += (
w_b07 * b07_norm + w_b8A * b8a_norm + w_b11 * b11_norm + w_b12 * b12_norm
)
var_sum += (
w_viewZen_norm * viewZen_norm
+ w_sunZen_norm * sunZen_norm
+ w_relAzim_norm * relAzim_norm
)
return tansig(var_sum)
@staticmethod
def _layer2(
cste, w_n1, neuron1, w_n2, neuron2, w_n3, neuron3, w_n4, neuron4, w_n5, neuron5
):
"""Define the second layer, which is a linear combination of the neurons => inputs of the activation function"""
return (
cste
+ w_n1 * neuron1
+ w_n2 * neuron2
+ w_n3 * neuron3
+ w_n4 * neuron4
+ w_n5 * neuron5
)
[docs] def apply_normalize(self):
#https://github.com/senbox-org/s2tbx/blob/master/s2tbx-biophysical/src/main/resources/auxdata/2_1/FCOVER/FCOVER_Normalisation
self.b03_norm = self._normalize(self.B03, 0, 0.253061520471542)
self.b04_norm = self._normalize(self.B04, 0, 0.290393577911328)
self.b05_norm = self._normalize(self.B05, 0, 0.305398915248555)
self.b06_norm = self._normalize(self.B06, 0.006637972542253, 0.608900395797889)
self.b07_norm = self._normalize(self.B07, 0.013972727018939, 0.753827384322927)
self.b8a_norm = self._normalize(self.B8A, 0.026690138082061, 0.782011770669178)
self.b11_norm = self._normalize(self.B11, 0.016388074192258, 0.493761397883092)
self.b12_norm = self._normalize(self.B12, 0, 0.493025984460231)
degToRad = math.pi / 180
funccos = np.vectorize(lambda t: math.cos(t))
self.viewZen_norm = self._normalize(
funccos(self.viewZenithMean * degToRad), 0.918595400582046, 1
)
self.sunZen_norm = self._normalize(
funccos(self.sunZenithAngles * degToRad),
0.342022871159208,
0.936206429175402,
)
self.relAzim_norm = funccos(
(self.sunAzimuthAngles - self.viewAzimuthMean) * degToRad
)
[docs] def get_lai(self):
"""Define biophysical vegetation index Leaf Area Index, computed from a trained neural network which has as inputs the metadata of sentinel2 images"""
n1 = self._neuron(
4.96238030555279, #bias of neuron 1 https://github.com/senbox-org/s2tbx/blob/master/s2tbx-biophysical/src/main/resources/auxdata/2_1/LAI/LAI_Weights_Layer1_Bias
-0.023406878966470, # neuron https://github.com/senbox-org/s2tbx/blob/master/s2tbx-biophysical/src/main/resources/auxdata/2_1/LAI/LAI_Weights_Layer1_Neurons
+0.921655164636366,
0.135576544080099,
-1.938331472397950,
-3.342495816122680,
0.902277648009576,
0.205363538258614,
-0.040607844721716,
-0.083196409727092,
0.260029270773809,
0.284761567218845,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n2 = self._neuron(
1.416008443981500,
-0.132555480856684,
-0.139574837333540,
-1.014606016898920,
-1.330890038649270,
0.031730624503341,
-1.433583541317050,
-0.959637898574699,
+1.133115706551000,
0.216603876541632,
0.410652303762839,
0.064760155543506,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n3 = self._neuron(
1.075897047213310,
0.086015977724868,
0.616648776881434,
0.678003876446556,
0.141102398644968,
-0.096682206883546,
-1.128832638862200,
0.302189102741375,
0.434494937299725,
-0.021903699490589,
-0.228492476802263,
-0.039460537589826,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n4 = self._neuron(
1.533988264655420,
-0.109366593670404,
-0.071046262972729,
+0.064582411478320,
2.906325236823160,
-0.673873108979163,
-3.838051868280840,
1.695979344531530,
0.046950296081713,
-0.049709652688365,
0.021829545430994,
0.057483827104091,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n5 = self._neuron(
3.024115930757230,
-0.089939416159969,
0.175395483106147,
-0.081847329172620,
2.219895367487790,
1.713873975136850,
0.713069186099534,
0.138970813499201,
-0.060771761518025,
0.124263341255473,
0.210086140404351,
-0.183878138700341,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
l2 = self._layer2(
1.096963107077220,
-1.500135489728730,
n1,
-0.096283269121503,
n2,
-0.194935930577094,
n3,
-0.352305895755591,
n4,
0.075107415847473,
n5,
)
return self._denormalize(l2, 0.000319182538301, 14.4675094548151)
[docs] def get_cab(self):
"""Define biochemical vegetation index Chloro a+b, computed from a trained neural network which has as inputs the metadata of sentinel2 images"""
n1 = self._neuron(
4.242299670155190,
0.400396555256580,
0.607936279259404,
0.137468650780226,
-2.955866573461640,
-3.186746687729570,
2.206800751246430,
-0.313784336139636,
+0.256063547510639,
-0.071613219805105,
0.510113504210111,
0.142813982138661,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n2 = self._neuron(
-0.259569088225796,
-0.250781102414872,
0.439086302920381,
-1.160590937522300,
-1.861935250269610,
0.981359868451638,
1.634230834254840,
-0.872527934645577,
0.448240475035072,
0.037078083501217,
0.030044189670404,
0.005956686619403,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n3 = self._neuron(
3.130392627338360,
0.552080132568747,
-0.502919673166901,
6.105041924966230,
-1.294386119140800,
-1.059956388352800,
-1.394092902418820,
0.324752732710706,
-1.758871822827680,
-0.036663679860328,
-0.183105291400739,
-0.038145312117381,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n4 = self._neuron(
0.774423577181620,
0.211591184882422,
-0.248788896074327,
0.887151598039092,
1.143675895571410,
-0.753968830338323,
-1.185456953076760,
0.541897860471577,
-0.252685834607768,
-0.023414901078143,
-0.046022503549557,
-0.006570284080657,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n5 = self._neuron(
2.584276648534610,
0.254790234231378,
-0.724968611431065,
0.731872806026834,
2.303453821021270,
-0.849907966921912,
-6.425315500537270,
2.238844558459030,
-0.199937574297990,
0.097303331714567,
0.334528254938326,
0.113075306591838,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
l2 = self._layer2(
0.463426463933822,
-0.352760040599190,
n1,
-0.603407399151276,
n2,
0.135099379384275,
n3,
-1.735673123851930,
n4,
-0.147546813318256,
n5,
)
return self._denormalize(l2, 0.007426692959872, 873.908222110306)
[docs] def get_fapar(self):
"""Define biophysical vegetation index Fraction of Absorbed Photosynthetically Active Radiation, computed from a trained neural network which has as inputs the metadata of sentinel2 images"""
n1 = self._neuron(
-0.887068364040280,
0.268714454733421,
-0.205473108029835,
0.281765694196018,
1.337443412255980,
0.390319212938497,
-3.612714342203350,
0.222530960987244,
0.821790549667255,
-0.093664567310731,
0.019290146147447,
0.037364446377188,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n2 = self._neuron(
0.320126471197199,
-0.248998054599707,
-0.571461305473124,
-0.369957603466673,
0.246031694650909,
0.332536215252841,
0.438269896208887,
0.819000551890450,
-0.934931499059310,
0.082716247651866,
-0.286978634108328,
-0.035890968351662,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n3 = self._neuron(
0.610523702500117,
-0.164063575315880,
-0.126303285737763,
-0.253670784366822,
-0.321162835049381,
0.067082287973580,
2.029832288655260,
-0.023141228827722,
-0.553176625657559,
0.059285451897783,
-0.034334454541432,
-0.031776704097009,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n4 = self._neuron(
-0.379156190833946,
0.130240753003835,
0.236781035723321,
0.131811664093253,
-0.250181799267664,
-0.011364149953286,
-1.857573214633520,
-0.146860751013916,
0.528008831372352,
-0.046230769098303,
-0.034509608392235,
0.031884395036004,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n5 = self._neuron(
1.353023396690570,
-0.029929946166941,
0.795804414040809,
0.348025317624568,
0.943567007518504,
-0.276341670431501,
-2.946594180142590,
0.289483073507500,
1.044006950440180,
-0.000413031960419,
0.403331114840215,
0.068427130526696,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
l2 = self._layer2(
-0.336431283973339,
2.126038811064490,
n1,
-0.632044932794919,
n2,
5.598995787206250,
n3,
1.770444140578970,
n4,
-0.267879583604849,
n5,
)
return self._denormalize(l2, 0.000153013463222, 0.977135096979553)
[docs] def get_fcover(self):
"""Define biophysical vegetation index Leaf Area Index, computed from a trained neural network which has as inputs the metadata of sentinel2 images"""
n1 = self._neuron(
-1.45261652206,
-0.156854264841,
0.124234528462,
+0.235625516229,
-1.8323910258,
-0.217188969888,
+5.06933958064,
-0.887578008155,
-1.0808468167,
-0.0323167041864,
-0.224476137359,
-0.195523962947,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n2 = self._neuron(
-1.70417477557,
-0.220824927842,
+1.28595395487,
+0.703139486363,
-1.34481216665,
-1.96881267559,
-1.45444681639,
+1.02737560043,
-0.12494641532,
+0.0802762437265,
-0.198705918577,
+0.108527100527,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n3 = self._neuron(
1.02168965849,
-0.409688743281,
+1.08858884766,
+0.36284522554,
+0.0369390509705,
-0.348012590003,
-2.0035261881,
+0.0410357601757,
+1.22373853174,
+-0.0124082778287,
-0.282223364524,
+0.0994993117557,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n4 = self._neuron(
-0.498002810205,
-0.188970957866,
-0.0358621840833,
+0.00551248528107,
+1.35391570802,
-0.739689896116,
-2.21719530107,
+0.313216124198,
+1.5020168915,
+1.21530490195,
-0.421938358618,
+1.48852484547,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n5 = self._neuron(
-3.88922154789,
+2.49293993709,
-4.40511331388,
-1.91062012624,
-0.703174115575,
-0.215104721138,
-0.972151494818,
-0.930752241278,
+1.2143441876,
-0.521665460192,
-0.445755955598,
+0.344111873777,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
l2 = self._layer2(
-0.0967998147811,
+0.23080586765,
n1,
-0.333655484884,
n2,
-0.499418292325,
n3,
+0.0472484396749,
n4,
-0.0798516540739,
n5,
)
return self._denormalize(l2, 0.000181230723879, 0.999638214715)
[docs] def get_cw(self):
"""Define biophysical vegetation index Fraction of Absorbed Photosynthetically Active Radiation, computed from a trained neural network which has as inputs the metadata of sentinel2 images"""
n1 = self._neuron(
-2.1064083686,
0.146378710426,
1.18979928187,
-0.906235139963,
-0.808337508767,
-0.97333491783,
-1.42591277646,
-0.00561253629588,
-0.634520356267,
-0.117226059989,
-0.0602700912102,
0.229407587132,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n2 = self._neuron(
-1.69022094794,
0.283319173374,
0.149342023041,
1.08480588387,
-0.138658791035,
-0.455759407329,
0.420571438078,
-1.7372949037,
-0.704286287226,
0.0190953782358,
-0.0393971316513,
-0.00750241581744,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n3 = self._neuron(
3.10117655255,
-0.197487427943,
-0.105460325978,
0.158347670681,
2.14912426654,
-0.970716842916,
-4.92725317909,
1.42034301781,
1.45316917226,
0.0227257053609,
0.269298650421,
0.0849047657715,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n4 = self._neuron(
-1.31231626496,
0.141405799763,
0.33386260328,
0.356218929123,
-0.545942267639,
0.0891043076856,
0.919298362929,
-1.8520892625,
-0.427539590779,
0.00791385646467,
0.0148333201478,
-0.00153786769736,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
n5 = self._neuron(
1.01131930348,
-0.186781083395,
-0.549163704901,
-0.181287638772,
0.96864043656,
-0.470442559117,
-1.24859725244,
2.67014942338,
0.49009062438,
-0.00144931939526,
0.00314829369692,
0.0206517883893,
self.b03_norm,
self.b04_norm,
self.b05_norm,
self.b06_norm,
self.b07_norm,
self.b8a_norm,
self.b11_norm,
self.b12_norm,
self.viewZen_norm,
self.sunZen_norm,
self.relAzim_norm,
)
l2 = self._layer2(
-0.197591709977,
-0.0775555890347,
n1,
-0.86411786119,
n2,
-0.199212415374,
n3,
1.98730461219,
n4,
0.458926743489,
n5,
)
return self._denormalize(l2, 3.85066859366e-06, 0.522417054645)
[docs]class VegetationIndicesS2(EOTask):
"""Define a class of vegetation indices, which are computed from the metadata of sentinel2 images extracted"""
def __init__(self, feature_name, mask_data=True):
self.feature_name = feature_name
self.mask_data = mask_data
[docs] def get_vegetation_indices(self):
"""Define vegetation indices which are simply ratio of spectral bands"""
self.NDVI = (self.B8A - self.B04) / (self.B8A + self.B04)
self.NDWI = (self.B8A - self.B11) / (self.B8A + self.B11)
self.GNDVI = (self.B8A - self.B03) / (self.B8A + self.B03)
self.NDVIre = (self.B8A - self.B05) / (self.B8A + self.B05)
[docs] def get_biophysical_parameters(self):
biopysicial_parameters = BiophysicalParameters(
self.B03,
self.B04,
self.B05,
self.B06,
self.B07,
self.B8A,
self.B11,
self.B12,
self.viewZenithMean,
self.sunZenithAngles,
self.viewAzimuthMean,
self.sunAzimuthAngles,
)
# Normalized bands
biopysicial_parameters.apply_normalize()
self.fapar = biopysicial_parameters.get_fapar()
self.LAI = biopysicial_parameters.get_lai()
self.CCC = biopysicial_parameters.get_cab()
self.Cab = self.CCC / self.LAI
self.CCW = biopysicial_parameters.get_cw()
self.CW = self.CCW / self.LAI
self.fcover = biopysicial_parameters.get_fcover()
[docs] def mask_pixels(self, eopatch):
bands_array = eopatch.data[self.feature_name]
valid_data_mask = (
eopatch.mask["VALID_DATA"] if self.mask_data else eopatch.mask["IS_DATA"]
)
if "polygon_mask" in list(eopatch.mask_timeless.keys()):
bands_array = np.ma.array(
bands_array,
dtype=np.float32,
mask=np.logical_or(
~valid_data_mask.astype(bool), np.isnan(bands_array)
),
fill_value=np.nan,
)
bands_array = bands_array.filled()
return bands_array
def execute(self, eopatch):
"""Add those vegeation indices to the eo-patch for futur use"""
bands_array = self.mask_pixels(eopatch)
illumination_array = eopatch.data["ILLUMINATION"]
# Raw data
self.B02 = bands_array[..., 0]
self.B03 = bands_array[..., 1]
self.B04 = bands_array[..., 2]
self.B05 = bands_array[..., 3]
self.B06 = bands_array[..., 4]
self.B07 = bands_array[..., 5]
self.B08 = bands_array[..., 6]
self.B8A = bands_array[..., 7]
self.B11 = bands_array[..., 8]
self.B12 = bands_array[..., 9]
self.viewZenithMean = illumination_array[..., 0]
self.sunZenithAngles = illumination_array[..., 1]
self.viewAzimuthMean = illumination_array[..., 2]
self.sunAzimuthAngles = illumination_array[..., 3]
self.get_vegetation_indices()
self.get_biophysical_parameters()
add_feat = AddFeatureTask((FeatureType.DATA, "fapar"))
add_feat.execute(eopatch=eopatch, data=self.fapar[..., np.newaxis])
add_feat = AddFeatureTask((FeatureType.DATA, "LAI"))
add_feat.execute(eopatch=eopatch, data=self.LAI[..., np.newaxis])
add_feat = AddFeatureTask((FeatureType.DATA, "CCC"))
add_feat.execute(eopatch=eopatch, data=self.CCC[..., np.newaxis])
add_feat = AddFeatureTask((FeatureType.DATA, "Cab"))
add_feat.execute(eopatch=eopatch, data=self.Cab[..., np.newaxis])
add_feat = AddFeatureTask((FeatureType.DATA, "CCW"))
add_feat.execute(eopatch=eopatch, data=self.CCW[..., np.newaxis])
add_feat = AddFeatureTask((FeatureType.DATA, "CW"))
add_feat.execute(eopatch=eopatch, data=self.CW[..., np.newaxis])
add_feat = AddFeatureTask((FeatureType.DATA, "fcover"))
add_feat.execute(eopatch=eopatch, data=self.fcover[..., np.newaxis])
add_NDVI = AddFeatureTask((FeatureType.DATA, "NDVI"))
add_NDVI.execute(eopatch=eopatch, data=self.NDVI[..., np.newaxis])
add_NDWI = AddFeatureTask((FeatureType.DATA, "NDWI"))
add_NDWI.execute(eopatch=eopatch, data=self.NDWI[..., np.newaxis])
add_GNDVI = AddFeatureTask((FeatureType.DATA, "GNDVI"))
add_GNDVI.execute(eopatch=eopatch, data=self.GNDVI[..., np.newaxis])
add_NDVIre = AddFeatureTask((FeatureType.DATA, "NDVIre"))
add_NDVIre.execute(eopatch=eopatch, data=self.NDVIre[..., np.newaxis])
if "ILLUMINATION" in eopatch.data.keys():
remove_task = RemoveFeatureTask((FeatureType.DATA, "ILLUMINATION"))
remove_task.execute(eopatch)
return eopatch
[docs]class EuclideanNorm(EOTask):
"""
The tasks calculates Euclidian Norm of all bands within an array:
norm = sqrt(sum_i Bi**2),
where Bi are the individual bands within user-specified feature array.
"""
def __init__(self, feature_name, in_feature_name):
self.feature_name = feature_name
self.in_feature_name = in_feature_name
def execute(self, eopatch):
arr = eopatch.data[self.in_feature_name]
norm = np.sqrt(np.sum(arr**2, axis=-1))
add_ecnorm = AddFeatureTask((FeatureType.DATA, self.feature_name))
return add_ecnorm.execute(eopatch=eopatch, data=norm[..., np.newaxis])
