-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy path03_get_landsatvi2nc.py
256 lines (204 loc) · 8.21 KB
/
03_get_landsatvi2nc.py
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
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
#%%
import os
import boto3
import rasterio as rio
from pystac_client import Client
import xarray as xr
import numpy as np
from pathlib import Path
import geopandas as gpd
from datetime import date, datetime
from utils import *
import matplotlib.pyplot as plt
#%%
run = 'f5'
farm_file = '/home/geodata/Clientes/0FARMS/FabioRicardi-Barreiras_BA/vetorial/CARs-Fazenda_Savana.gpkg'
farm_file = '/home/geodata/Clientes/0FARMS/MG-3102605-B4D344DBFD874F44906FCC0A5E0DCE36/CAR.gpkg'
layer = 'car'
folderout = '/home/geodata/Clientes/0FARMS/FabioRicardi-Barreiras_BA/nc/'
folderout = '/home/geodata/Clientes/0FARMS/MG-3102605-B4D344DBFD874F44906FCC0A5E0DCE36/nc/'
# Satellite imagery query params
today = date.today()
datetime_rangefull = str(f"2015-06-20/{str(today)}") #break 2017
#datetime_rangefull = str(f"1985-06-20/2013-06-20") #break 2017
max_cloud = 100
#bucketname = 'sanca'
satellite = 'Landsat'
platforms = ["LANDSAT_5","LANDSAT_8", "LANDSAT_9"] #"LANDSAT_7",
folder_nc = f'{folderout}{run}/'
print(folder_nc)
# Set environment and create AWS Session
os.environ['CURL_CA_BUNDLE'] = '/etc/ssl/certs/ca-certificates.crt'
os.environ['AWS_REQUEST_PAYER'] = 'requester'
print("Creating AWS Session")
aws_session = rio.session.AWSSession(boto3.Session(), requester_pays=True)
print(aws_session)
# open Farm
try:
farm = gpd.read_file(farm_file, layer=layer)
except:
farm = gpd.read_file(farm_file)
bbox = get_bbox(farm)
# %% GET COLLECTION AND PARAMETERS
URL = 'https://landsatlook.usgs.gov/stac-server'
cat = Client.open(URL)
collection_id = 'landsat-c2l2-sr'
collection = cat.get_collection(collection_id)
print(collection)
query_params = {
"eo:cloud_cover": {"lt": max_cloud},
"platform": {"in": platforms},
"landsat:collection_category": { "in": ['T1']}
}
assets = ['red', 'blue', 'nir08', 'swir16'] #'green', 'qa_pixel'
y0 = datetime.strptime(datetime_rangefull.split('/')[0],'%Y-%m-%d').year
yf = datetime.strptime(datetime_rangefull.split('/')[1],'%Y-%m-%d').year
m0 = str(datetime.strptime(datetime_rangefull.split('/')[0],'%Y-%m-%d').month).zfill(2)
mf = str(datetime.strptime(datetime_rangefull.split('/')[1],'%Y-%m-%d').month).zfill(2)
d0 = str(datetime.strptime(datetime_rangefull.split('/')[0],'%Y-%m-%d').day+1).zfill(2)
df = str(datetime.strptime(datetime_rangefull.split('/')[1],'%Y-%m-%d').day-2).zfill(2)
if int(d0)>31: d0 = '30'
if int(df)<1: df = '01'
# %%
Path(folder_nc).mkdir( parents = True, exist_ok = True)
for ano in range(y0,yf):
'''
TODO falta um globals aqui pra player com os dados
'''
d0_ = str(int(d0)-2).zfill(2)
datetime_range = f'{ano}-{m0}-{d0}/{ano+1}-{m0}-{d0_}'
if yf == (ano+1):
datetime_range = f'{ano}-{m0}-{d0}/{ano+1}-{mf}-{df}'
print(ano, datetime_range)
#datetime_range = '2013-06-20/2024-05-08' SE QUISER FAZER A LOUCURA DE MANDAR TUDO DE UMA VEZ
datetime_range_name = datetime_range.replace('/','_')
try:
ds = get_cube(datetime_range,
cat,
collection_id,
bbox,
query_params,
aws_session, assets)
ds = dropper( ds , sat = satellite )
# é aqui que a filtragem acontece
ds2 = ds.copy()
# valores extremos
for asset in assets:
ds2[asset] = xr.where(ds2[asset] > 40000, np.nan, ds2[asset])
ds2[asset] = xr.where(ds2[asset] < 5000, np.nan, ds2[asset])
print(asset)
quantiles = [0.01,0.1,0.25,0.5,0.75,0.9,0.99]
print(np.nanquantile(ds2[asset],quantiles))
ds2[asset] = xr.where(ds2[asset] < np.nanquantile(ds2[asset],[0.01]), np.nan, ds2[asset])
ds2[asset] = xr.where(ds2[asset] > np.nanquantile(ds2[asset],[0.99]), np.nan, ds2[asset])
if asset == 'blue':
ds2[asset] = xr.where(ds2[asset] > np.nanquantile(ds2[asset],[0.84]), np.nan, ds2[asset])
# interpolate_na
ds2 = ds2.chunk(dict(time=-1))
ds2 = ds2.interpolate_na(dim="time",
method='linear',
use_coordinate=True,
)
# rolling
w = 3
ds2 = ds2.rolling(time=w, center=True).mean(skipna=True)
# # REPROJECTION
print(f'reprojecting cube for {datetime_range}')
ds2 = ds2.rio.write_crs('epsg:4326')
ds2 = ds2.rio.reproject('EPSG:4326')
ds2 = ds2.rename({'x': 'longitude','y': 'latitude'})
print('reprojecting... done')
ds2.to_netcdf(f'{folder_nc}/raw_{datetime_range_name}.nc')
print(f'>> {folder_nc}/raw_{datetime_range_name}.nc')
# CALCULATE indices
ndvi = NDVI( ds2 )
bsi= BSI( ds2 )
# get dates to save exact name
t0 = str(ndvi.time[0].values).split('T')[0]
t1 = str(ndvi.time[-1].values).split('T')[0]
n = len(ndvi.time)
try:
ndvi = dropper(ndvi, 'Landsat')
bsi = dropper(bsi, 'Landsat')
except:
print('not dropping, probably wont save')
# save nc
ndvi.to_netcdf(f'{folder_nc}/ndvi_{t0}-{t1}_{n}.nc')
print('> ndvi saved')
bsi.to_netcdf(f'{folder_nc}/bsi_{t0}-{t1}_{n}.nc')
print('> bsi saved')
except:
print(f'{datetime_range} FAILED')
# %% PLOTTING STUFF
# # ndvi.clip(0,1000)
# # # %% TEMOS UM ds
# #%% a copy
# ds2 = ds.copy(deep=True)
# for asset in assets:
# plt.hist(np.ravel(ds2[asset].values), bins = 100)
# plt.title(asset); plt.grid();plt.show(); plt.close()
# ts = ds2.sel(longitude = -46.64329, latitude = -22.07701, method = 'nearest')
# for asset in assets:
# plt.plot(ts[asset].values, label = asset)
# #%% filta os principais outliers
# ds2 = ds.copy(deep=True)
# ds2 = xr.where(ds2 > 40000, np.nan, ds2)
# ds2 = xr.where(ds2 < 5000, np.nan, ds2)
# for asset in assets:
# quantiles = [0.01,0.1,0.25,0.5,0.75,0.9,0.99]
# print(quantiles)
# print(np.nanquantile(ds2[asset],quantiles))
# ds2[asset] = xr.where(ds2[asset] < np.nanquantile(ds2[asset],[0.01]), np.nan, ds2[asset])
# ds2[asset] = xr.where(ds2[asset] > np.nanquantile(ds2[asset],[0.99]), np.nan, ds2[asset])
# if asset == 'blue':
# ds2[asset] = xr.where(ds2[asset] > np.nanquantile(ds2[asset],[0.76]), np.nan, ds2[asset])
# #ds2[asset] = xr.where(ds2[asset] < np.nanquantile(ds2[asset],[0.01]), np.nan, ds2[asset])
# plt.hist((np).ravel(ds2[asset].values), bins = 100)
# plt.title(asset); plt.grid();plt.show(); plt.close()
# # farm.centroid
# ts2 = ds2.sel(longitude = -46.64329, latitude = -22.07701, method = 'nearest')
# for asset in assets:
# plt.plot(ts2[asset].values, label = asset)
# plt.legend()
# plt.grid()
# # %%
# #f -> Interpolate NaN
# #ds3 = ds2.copy()
# method = 'linear'
# ds3 = ds2.interpolate_na(dim="time",
# method=method,
# use_coordinate=True,
# ) # pchip # limit = 7, use_coordinate=True,
# ts3 = ds3.sel(longitude = -46.64329, latitude = -22.07701, method = 'nearest')
# for asset in assets:
# plt.plot(ts3[asset].values, label = asset)
# plt.grid();plt.show();plt.close()
# for asset in assets:
# plt.hist(np.ravel(ds3[asset].values), bins = 100)
# plt.title(asset); plt.grid();plt.show(); plt.close()
# # %%
# # f -> rolling
# w = 3
# ds4 = ds3.rolling(time=w, center=True).mean(skipna=True)#savgol_filter, window=w, polyorder=2
# ts4 = ds4.sel(longitude = -46.64329, latitude = -22.07701, method = 'nearest')
# for asset in assets:
# plt.plot(ts4[asset].values, label = asset)
# plt.grid();plt.show();plt.close()
# for asset in assets:
# plt.hist(np.ravel(ds4[asset].values), bins = 100)
# plt.title(asset); plt.grid();plt.show(); plt.close()
# # %%
# for asset in assets:
# plt.hist(np.ravel(ds4[asset].values), bins = 100)
# plt.title(asset)
# plt.show(); plt.close()
# ts2 = ds3.sel(longitude = -46.64329, latitude = -22.07701, method = 'nearest')
# for asset in assets:
# plt.plot(ts2[asset].values, label = asset)
# plt.legend()
# plt.grid()
# for asset in assets:
# plt.hist(np.ravel(ds4[asset].values), bins = 100)
# plt.title(asset)
# plt.title(asset); plt.grid();plt.show(); plt.close()
# %%