def download_error_image(img_date,geo_img,img_id,username,password):
'''
After read error file(image_error.txt) you can get image info which you failed from COG Sentinel-2, you can use this info with this function
if you have more than 1 image, you can download with for loop.
You can find img_date, geo_img and img_id information in image_error.txt file.
api,target_image_id=download_error_image(img_date,geo_img,img_id,username,password)
api.download(target_image_id,directory_path='.')
api.download('7be30c50-31fc-48c4-ab45-fddea9be7877',directory_path='.')
if you get error like >> Product 7be30c50-31fc-48c4-ab45-fddea9be7877 is not online. Triggering retrieval from long term archive.
Go to https://sentinelsat.readthedocs.io/en/stable/api.html#lta-products
username and password should be string
'''
api = SentinelAPI(username, password, 'https://scihub.copernicus.eu/dhus')
day_before =img_date- datetime.timedelta(days=1)
day_after =img_date + datetime.timedelta(days=1)
footprint = geojson_to_wkt(geo_img)
products = api.query(footprint,
#date = ('20181219', date(2018, 12, 29)),
date=(day_before,day_after),
platformname = 'Sentinel-2',
)
sat_df=api.to_geodataframe(products)
result=sat_df.loc[sat_df['title']==img_id]
return api,result.index.values[0]
def get_products(login_json, coordinates, date_start, date_end, download_dir):
with open(login_json, 'r') as fp:
LOGIN_INFO = json.load(fp)
USER_NAME, PASSWORD = list(LOGIN_INFO.values())
# connect to the API
api = SentinelAPI(USER_NAME, PASSWORD, 'https://scihub.copernicus.eu/dhus')
# define a map polygon
geojson = Polygon(coordinates=coordinates)
# search by polygon, time, and Hub query keywords
footprint = geojson_to_wkt(geojson)
dates = (date_start, date_end) # (date(2018, 4, 1), date(2018, 4, 11))
# June to July maps
products = api.query(
footprint,
date=dates,
platformname='Sentinel-2',
# producttype='S2MSI2A',
area_relation='Intersects', # area of interest is inside footprint
cloudcoverpercentage=(0, 40))
# download all results from the search
api.download_all(products, directory_path=download_dir)
# product_id = list(products.keys())[0]
# api.download(id=product_id, directory_path=download_dir)
# GeoPandas GeoDataFrame with the metadata of the scenes and the footprints as geometries
return api.to_geodataframe(products)
def test_to_geopandas():
api = SentinelAPI(**_api_auth)
products = api.query(get_coordinates('tests/map.geojson'),
"20151219",
"20151228",
platformname="Sentinel-2")
gdf = api.to_geodataframe(products)
api = SentinelAPI(user, password, 'https://scihub.copernicus.eu/dhus')
#Setting the search footprint using a Geojson file
footprint = geojson_to_wkt(read_geojson(AOI))
#Setting the search query parameters
products = api.query(footprint,
date = ('20200101', '20200331'),
platformname = 'Sentinel-2',
processinglevel = 'Level-2A',
cloudcoverpercentage = (0, 20))
#Printing the number of products found
print("The number of products found is: {} " .format (len(products)))
#Creating a table with all the product search results
products_table = api.to_geodataframe(products)
#This part downloads the product(s) in the same folder where your code is located
## OPTION 1: Download single product
api.download('df132697-676e-43ce-b7bd-45211696119f')
## OPTION 2:Download all products
download_list = []
for index, row in products_table.iterrows():
download_list.append (row['title'])
print ("The following products will be downloaded: {}" . format (download_list))
api.download (row['uuid'])
def download_images(save_imgs, save_rgb, save_tiles, unet_weights, unet_clouds,
class_path, class_clouds, poly_path, percentiles_forest,
percentiles_clouds, boundsdata):
# connect to the API
user = '******'
password = '******'
api = SentinelAPI(user, password, 'https://scihub.copernicus.eu/dhus')
# search by polygon
footprint = geojson_to_wkt(read_geojson(boundsdata))
# search for the images
products = api.query(
footprint,
date=(["NOW-30DAYS", "NOW"]),
area_relation='IsWithin',
platformname='Sentinel-2',
processinglevel='Level-2A',
#cloudcoverpercentage = (0, 20)
)
print(len(products))
table_names = api.to_geodataframe(products)
uuid_names = table_names['uuid'] # names to download
name_zip = table_names['title'] #title of image
extracted_name = table_names['filename'] # name of folder .SAFE
k = 0
# download images
for fname in uuid_names:
file_dir = save_imgs + '/' + extracted_name[k]
if os.path.isdir(file_dir) is False:
retval = os.getcwd()
os.chdir(save_imgs)
print("Downloading data...")
api.get_product_odata(fname)
api.download(fname)
os.chdir(retval) # return to previous directory
path_zip_name = save_imgs + '/' + name_zip[k] + '.zip'
while not os.path.exists(path_zip_name):
time.sleep(1)
if os.path.isfile(path_zip_name):
# extract files
zip_ref = zipfile.ZipFile(path_zip_name, 'r')
zip_ref.extractall(save_imgs)
zip_ref.close()
os.remove(path_zip_name) # remove .zip file
print("%s has been removed successfully" % name_zip[k])
path_to_folder = save_imgs + '/' + extracted_name[
k] + '/GRANULE/'
# calls the rgb_tiles function
dir_save_tiles = save_tiles + '/' + name_zip[k]
if os.path.isdir(dir_save_tiles) is False:
print('Creating RGB tiles')
os.mkdir(dir_save_tiles)
rgb_tiles(path_to_folder, save_rgb, dir_save_tiles,
name_zip[k])
# calls the application() Unet function
save_class_path = class_path + '/' + name_zip[k]
if os.path.isdir(save_class_path) is False:
print('Applying UNet')
os.mkdir(save_class_path)
application(dir_save_tiles,
unet_weights,
save_class_path,
percentiles_forest,
clouds=0)
# merge predicted tiles into one GeoTiff
join_tiles(save_class_path, class_path, path_to_folder)
print("Tiles merged!")
save_class_clouds = class_clouds + '/' + name_zip[k]
if os.path.isdir(save_class_clouds) is False:
print('Applying UNet clouds')
os.mkdir(save_class_clouds)
application(dir_save_tiles,
unet_clouds,
save_class_clouds,
percentiles_clouds,
clouds=1)
# merge the clouds predicted tiles into one GeoTiff
join_tiles(save_class_clouds, class_clouds, path_to_folder)
print("Clouds tiles merged!")
# polygons evalutation
print("Polygons evaluation")
polygons(name_zip[k],
class_path,
class_clouds,
path_to_folder,
save_class_path,
save_imgs,
poly_path,
time_spaced=None)
k = k + 1
else:
raise ValueError("%s isn't a file!" % path_zip_name)
else:
path_to_folder = save_imgs + '/' + extracted_name[k] + '/GRANULE/'
# calls the rgb_tiles function
dir_save_tiles = save_tiles + '/' + name_zip[k]
if os.path.isdir(dir_save_tiles) is False:
print('Creating RGB tiles')
os.mkdir(dir_save_tiles)
rgb_tiles(path_to_folder, save_rgb, dir_save_tiles,
name_zip[k])
# calls the application() Unet function
save_class_path = class_path + '/' + name_zip[k]
if os.path.isdir(save_class_path) is False:
print('Applying UNet')
os.mkdir(save_class_path)
application(dir_save_tiles,
unet_weights,
save_class_path,
percentiles_forest,
clouds=0)
# merge predicted tiles into one GeoTiff
join_tiles(save_class_path, class_path, path_to_folder)
print("Tiles merged!")
save_class_clouds = class_clouds + '/' + name_zip[k]
if os.path.isdir(save_class_clouds) is False:
print('Applying UNet clouds')
os.mkdir(save_class_clouds)
application(dir_save_tiles,
unet_clouds,
save_class_clouds,
percentiles_clouds,
clouds=1)
# merge the clouds predicted tiles into one GeoTiff
join_tiles(save_class_clouds, class_clouds, path_to_folder)
print("Clouds tiles merged!")
# polygons evalutation
print("Polygons evaluation")
polygons(name_zip[k],
class_path,
class_clouds,
path_to_folder,
save_class_path,
save_imgs,
poly_path,
time_spaced=None)
k = k + 1
return
'31TCG', '30TYN', '30TYM', '30TYL', '30TYK', '30TXN', '30TXM', '30TXL',
'30TXK', '30TWM', '30TWL'
]
#first search l2a data
gdfl2 = gpd.GeoDataFrame()
product = ['S2MSI2A', 'S2MSI2AP']
for i in tile:
for j in product:
productstmp = api.query(filename='*%s*' % i,
beginPosition=('20180101', '20181231'),
producttype='%s' % j,
platformname='Sentinel-2')
gdftmp = api.to_geodataframe(productstmp)
gdfl2 = pd.concat([gdfl2, gdftmp], sort=True)
gdfl2.to_csv('SPN_l2a_18.csv')
#then search l1c data
gdfl1c = gpd.GeoDataFrame()
for i in tile:
productstmp = api.query(filename='*%s*' % i,
beginPosition=('20160101', '20171231'),
producttype='S2MSI1C',
platformname='Sentinel-2')
gdftmp = api.to_geodataframe(productstmp)
