Entre no USGS Earth Explorer e baixe imagens do Landsat com Python

9

Quero fazer login no site do USGS Earth Explorer em http://earthexplorer.usgs.gov/ com Python e baixar imagens da coleção do Landsat Archive apenas fornecendo nome de usuário, senha, caminho / linha (ou lat / long), data e limite de cobertura de nuvens. Eu já tentei isso: https://github.com/olivierhagolle/LANDSAT-Download/wiki Não funcionou. Em seguida, analisei o seguinte: https://github.com/developmentseed/landsat-util Mas não pude adicionar a ferramenta devido a restrições de trabalho, por isso não posso usá-la.

Assim. Eu tenho uma experiência limitada em Python e nunca acessei um site com ele antes. Estou usando o Python 2.7.8 no Windows 7 e tentando acessar o site pelo Chrome (versão 51.0.2704.106 m, se isso ajudar)

Alguma idéia para um script simples para acessar o site, fazer login (já tenho credenciais de login), procurar imagens e fazer o download para uma pasta? Parece que algo que o Python deve ser capaz de fazer, mas todas as soluções on-line parecem avançadas demais para mim. Eu só preciso de um script independente que possa ser usado repetidamente com diferentes coordenadas e datas.

MattS
fonte
O landsat-util é apenas um script python. Se você não pode adicionar dependências, pode precisar dizer o que já possui. Você pode usar o mapbox.github.io/usgs (também github.com/mapbox/usgs )?
21716 BradHards
1
Observe que dizer "Não deu certo" é uma maneira infalível de fazer as pessoas ignorarem sua pergunta. "Não funcionou" não nos diz nada de útil. Se você quiser ajudar a fazê-lo funcionar, edite sua pergunta com mais informações.
precisa saber é o seguinte
@ Lucas: como em, eu executei e não cumpriu a tarefa. Não fez login no site nem baixou os arquivos. Não tenho certeza de como isso pode ser mal interpretado. O script faz. não. trabalhos. Então, eu estou procurando métodos alternativos.
Matts
1
Sendo apenas um advogado do diabo, o que há de errado com o aplicativo BDA para download em massa?
Hornbydd
@Ornbydd: não há nada errado com o BDA. Eu uso isso o tempo todo. Eu acho que essa é minha única opção real. Não existe um bom caminho pitônico.
Matts

Respostas:

1

Eu recebi o código MUITO bagunçado que eu uso para baixar todos os dados do Landsat definindo os parâmetros Path e Row. Além disso, existem algumas funções personalizadas e do arco-íris que não são necessárias para você. Você pode limpar este código para adaptá-lo para seus propósitos (os comentários estão em russo):

# -*- coding: utf-8 -*-
import os
import sys
import shutil
import subprocess
import traceback
import time
import requests
from pprint import pprint

# выключить предупреждение
requests.packages.urllib3.disable_warnings()
# -----------------------------------------------------------
def find_scenes_in_DB(FC_in_GDB, field_names_list = ['ID', 'ARCHIVE']):
    with arcpy.da.SearchCursor(FC_in_GDB, field_names_list) as DB_cursor:
        uniq_sat_name = sorted({row[0] for row in DB_cursor})
        DB_cursor.reset()
        uniq_sat_archive = sorted({row[1] for row in DB_cursor})
    uniq_sat_name += uniq_sat_archive  
    return uniq_sat_name
# -----------------------------------------------------------
# замеры времени
def Time_now():
    return time.time()
# -----------------------------------------------------------
def Time_elapsed(time_start, time_end):
    return time_end - time_start
# -----------------------------------------------------------
# сгенерировать текст запроса (URL) и вернуть его в виде текста

def generate_json_request(
        request_code,
        json_request_content,
        http_service_endpoint=r'https://earthexplorer.usgs.gov/inventory/json/v/1.4.0/'
):
    return str(http_service_endpoint + request_code + r'?jsonRequest=' + json_request_content)
# -----------------------------------------------------------
# авторизироваться на сайте и получить токен доступа
def login(username, password, catalogId, authType = 'EROS'):
    URL = generate_json_request(
        request_code='login',
        json_request_content='{' +
        '"username":"'    + username  +
        '","password":"'  + password  +
        '","authType":"'  + authType  + 
        '","catalogId":"' + catalogId +
        '"}'
    )

    # послать POST запрос
    answer = requests.post(URL)

    # проверка прошёл ли запрос
    check_answer = (answer.status_code, answer.reason)

    if check_answer != (200, 'OK'):
        print u'Ошибка! Сайт не доступен.', check_answer, u'Ссылка:'
        print URL[:150], '...'
        return None

    else:
        # если ошибка (авторизация провалена)
        if answer.json()['errorCode'] is not None:
            print u'Ошибка!', answer.json()['errorCode'] + ':', answer.json()['error']

            # else:
            #     print u'Авторизация успешна! Токен:', answer.json()['data']
        return answer.json()['data']

# удалить токен авторизации
def logout(apiKey):
    if apiKey is not None:

        URL = generate_json_request(
            request_code='logout',
            json_request_content='{"apiKey":"' + apiKey + '"}'
        )

        # послать POST запрос
        answer = requests.post(URL)

        # проверка прошёл ли запрос
        check_answer = (answer.status_code, answer.reason)

        if check_answer != (200, 'OK'):
            print u'Ошибка! Сайт не доступен.', check_answer, u'Ссылка:'
            print URL[:150], '...'

        else:
            # если ошибка
            if answer.json()['errorCode'] is not None:
                #            print u'Ошибка!', answer.json()['errorCode'] + ':', answer.json()['error']
                return None

            else:
                #            if  answer.json()['data'] == False:
                #                print u'Токен НЕ удалён.'
                #                pprint(answer.json())
                #            else:
                #                print u'Токен удалён.'
                return answer.json()['data']

def datasetfields(datasetName, apiKey, node='EE'):
    URL = generate_json_request(
        request_code='datasetfields',
        json_request_content='{"apiKey":"' + apiKey + '", "node": "' + node + '", "datasetName": "' + datasetName + '"}'
    )

    # послать POST запрос
    answer = requests.post(URL)

    # проверка прошёл ли запрос
    check_answer = (answer.status_code, answer.reason)

    if check_answer != (200, 'OK'):
        print u'Ошибка! Сайт не доступен.', check_answer, u'Ссылка:'
        print URL[:150], '...'

    else:
        # если ошибка
        if answer.json()['errorCode'] is not None:
            print u'Ошибка!', answer.json()['errorCode'] + ':', answer.json()['error']
            return None
        else:
            return answer.json()

# сформировать критерий поиска снимков на основе списка Path Row
def additionalCriteria(list_Path_Row, datasetName):
    additionalCriteria = '{"filterType":"or", "childFilters":['
    # первая итерация?
    first_time = True

    for x in datasetfields(datasetName, apiKey).get('data'):
        if x.get('name') == 'WRS Path': Path_fieldId = x.get('fieldId')
        if x.get('name') == 'WRS Row':   Row_fieldId = x.get('fieldId')

    for PR in list_Path_Row:
        Path = str(PR[0])
        Row = str(PR[1])

        filter_Path = '{"filterType":"value", "fieldId":' + str(Path_fieldId) + ', "value":"' + Path + '", "operand":"="}'  # "fieldId":10036 - PATH
        filter_Row = '{"filterType":"value", "fieldId":' + str(Row_fieldId) + ', "value":"' + Row + '", "operand":"="}'  # "fieldId":10038 - ROW
        filter_And = '{"filterType":"and", "childFilters":[' + filter_Path + ',' + filter_Row + ']}'

        # если первая - не добавляем запятую в начало
        if first_time == True:
            additionalCriteria += filter_And
            first_time = False
        else:
            additionalCriteria += ' ,' + filter_And

    additionalCriteria += ']}'
    return additionalCriteria

# поиск снимков
def search_scenes(datasetName, list_Path_Row, apiKey,
                  startDate='2016-05-01T00:00:00Z',
                  endDate='2100-01-01T00:00:00Z',
                  months=[1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12],
                  maxResults=10,
                  ):
    URL = generate_json_request(
        request_code='search',
        json_request_content='{"datasetName":"' + datasetName + '"' + \
                             ',"temporalFilter":{"dateField": "search_date","startDate":"' + startDate + '","endDate":"' + endDate + '"}' + \
                             ',"months":' + str(months) + \
                             ',"maxResults":' + str(maxResults) + \
                             ',"additionalCriteria":' + additionalCriteria(list_Path_Row, datasetName) + \
                             ',"node":"EE","apiKey":"' + apiKey + '"}'
    )

    # послать POST запрос
    answer = requests.post(URL)

    # проверка прошёл ли запрос
    check_answer = (answer.status_code, answer.reason)

    if check_answer != (200, 'OK'):
        print u'Ошибка! Сайт не доступен.', check_answer, u'Ссылка:'
        print URL[:150], '...'
        print
        print URL

    else:
        # если ошибка
        if answer.json()['errorCode'] is not None:
            print u'Ошибка!', answer.json()['errorCode'] + ':', answer.json()['error']
            return None
        else:
            if answer.json()['data'] in (None, ''):
                print u'Снимков не найдено.'
            else:
                number_of_scenes = answer.json()['data']['totalHits']

            dwnld_URL_list = []
            for scene in answer.json()['data']['results']:
                dwnld_URL_list.append(scene['entityId'])

            return dwnld_URL_list

# вернуть список URL для прямой закачки снимков
def get_download_list(datasetName, entityIds, apiKey, node='EE', products='["STANDARD"]'):
    entityIds = str(['"' + x.encode('UTF8') + '"' for x in entityIds]).replace("'", '')

    URL = generate_json_request(
        request_code='download',
        json_request_content='{' + \
                             '"datasetName": "' + datasetName + '",' + \
                             '"apiKey": "' + apiKey + '",' + \
                             '"node": "' + node + '",' + \
                             '"entityIds": ' + str(entityIds) + ',' + \
                             '"products": ' + products + \
                             '}'
    )

    # послать POST запрос
    answer = requests.post(URL)

    # проверка прошёл ли запрос
    check_answer = (answer.status_code, answer.reason)

    if check_answer != (200, 'OK'):
        print u'Ошибка! Сайт не доступен.', check_answer, u'Ссылка:'
        print URL[:150], '...'
        return None

    else:
        # если ошибка (авторизация провалена)
        if answer.json()['errorCode'] is not None:
            print u'Ошибка!', answer.json()['errorCode'] + ':', answer.json()['error']
        return answer.json()['data']

# делаем список уже закаченных файлов
def list_of_scenes_in_archive(archive_path):
    # archive_path = archive_path.decode('utf-8')
    loaded = []
    dirList = os.listdir(archive_path)
    for fname in dirList:
        fname = fname[:fname.rfind('.')]  # удаляются 4 последних знака (в моём случае это ".rar")
        loaded.append(fname)
    return loaded

# скачать и сохранить файл
def download_file(url, file_path):
    r = requests.get(url, timeout=120, stream=True)
    with open(file_path, 'wb') as f:
        for chunk in r.iter_content(chunk_size=1024):
            f.write(chunk)
    return file_path

# тестировать файл формата tar.gz
def test_gz_archive(archive_path, GZIP_path):
    coding = 'cp1251'
    GZIP_path = GZIP_path.encode(coding)
    archive_path = archive_path.encode(coding)
    cmd_list = [GZIP_path, '-t', archive_path]
    # запуск консоли и выполнение команды в скрытном режиме с выводом
    subprocess.check_output(cmd_list, stderr=subprocess.STDOUT, shell=True)

def main(username, password, catalogId, datasets, WRS_2_Path_Row_list, WRS_1_Path_Row_list, archive_path, temp_dwnld_folder, FC_in_GDB):
    while True:

        try:
            print time.strftime("%d.%m.%Y %H:%M:%S"),
            global apiKey  # глобальная переменная!!!
            apiKey = login(username, password, catalogId)  # авторизация

            # список уже закаченных снимков
            loaded = list_of_scenes_in_archive(archive_path)
            print u"Всего файлов в локальном хранилище:", len(loaded)

            loaded_2 = find_scenes_in_DB(FC_in_GDB)
            loaded += loaded_2

            for dataset in datasets:
                # print '---------------------------'
                print u'Набор данных:', dataset

                if dataset in 'LANDSAT_MSS':
                    list_Path_Row = WRS_1_Path_Row_list
                else:
                    list_Path_Row = WRS_2_Path_Row_list

                # делаем запросы на каждый отдельный Path-Row, чтобы не превысить ограничение по длине URL
                for path_row in list_Path_Row:

                    entityIds = search_scenes(datasetName=dataset,
                                              list_Path_Row=[path_row],
                                              apiKey=apiKey,
                                              maxResults=50000)

                    # список сцен которых нет в архиве
                    dwnld_list = list(set(entityIds) - set(loaded))

                    if len(dwnld_list) > 0:
                        print time.strftime("%d.%m.%Y %H:%M:%S"),
                        print u'Path Row:', str(path_row[0]), str(path_row[1]) + '.',
                        print u'Всего снимков:', str(len(entityIds)) + '.',
                        print u'Новых:', str(len(dwnld_list)) + '.',
                        print u'Закачка файлов:'

                        # получаем сслыку на закачку для каждой отдельной сцены
                        for (scene_num, dwnld_scene) in enumerate(dwnld_list):

                            DICTs_list = get_download_list(dataset, [dwnld_scene], apiKey, node='EE',
                                                          products='["STANDARD"]')

                            # скачать снимки
                            for DICT in DICTs_list:
                                print time.strftime("%d.%m.%Y %H:%M:%S"), u'Cнимок:',
                                URL = DICT[u'url']

                                # формирование пути
                                scene_product_ID = URL[URL.rfind('/') + 1:URL.rfind('.tar.gz?')]   # название снимка по-новому
                                scene_ID = DICT[u'entityId'] # название снимка по-старому

                                scene_path = os.path.join(temp_dwnld_folder, scene_product_ID + u'.tar.gz')

                                print str(scene_num + 1), scene_product_ID

                                time_start = Time_now()  # текущее время (начало закачки)

                                try:
                                    ##                                    print
                                    ##                                    print '-'*10
                                    ##                                    print URL
                                    ##                                    print scene_path
                                    real_file_len = requests.head(URL).headers[
                                        'content-length']  # размер файла на сервере
                                    download_file(URL, scene_path)  # закачка
                                    dwnld_file_len = str(long(os.path.getsize(scene_path)))  # размер закаченного файла                                    

                                    if os.path.isfile(scene_path):
                                        if real_file_len != dwnld_file_len:                                            
                                            os.remove(scene_path)
                                            print u'<= DEL, файл не докачался'

                                    # print real_file_len
                                    # print dwnld_file_len
                                    # print real_file_len == dwnld_file_len
                                    # print '-'*10

                                    # проверка архива
                                    # test_gz_archive(scene_path, GZIP_path = ur'G:\Install\GZIP\gzip.exe')

                                    if scene_num + 1 != len(dwnld_list):  # вывести запятую
                                        print ',',

                                # если ошибка при закачке - удалить недокаченный файл
                                except:
                                    print u'<= DEL,',
                                    if os.path.isfile(scene_path):
                                        os.remove(scene_path)

                                    traceback.print_exc()  # напечатать ошибку

                                time_end = Time_now()  # текущее время (конец закачки)

                                # если закачка была менее или равно 60 минут назад
                                if Time_elapsed(time_start, time_end) <= 60 * 60:
                                    logout(apiKey)
                                    apiKey = login(username, password, catalogId)  # авторизация

                                elif Time_elapsed(time_start, time_end) > 60 * 60:
                                    apiKey = login(username, password, catalogId)  # авторизация

            # если авторизация была успешной - удалить токен авторизации
            if apiKey is not None: logout(apiKey)

        # если в процессе выполнения была ошибка
        except:
            try:
                logout(apiKey)  # удалить токен авторизации
            except:
                pass
            print
            traceback.print_exc()  # напечатать ошибку

        print time.strftime("%d.%m.%Y %H:%M:%S") + u' =========== Повтор через час ===========\n'
        time.sleep(3600)


if __name__ == "__main__":

    WRS_2_Path_Row_list=[
        [165, 14], [166, 14],
    ]

    WRS_1_Path_Row_list=[
        [182, 13], [176, 14],
    ]


    main(
        username='username',
        password='password',
        catalogId = "EE",  # возможные параметры catalogId: "EE" "GLOVIS" "HDDS" "LPCS"

        datasets=[
            'LANDSAT_8_C1',
##            'LANDSAT_8_PREWRS',
##            'LANDSAT_ETM_C1',
##            'LANDSAT_TM_C1',
##            'LANDSAT_MSS',
        ],
        WRS_2_Path_Row_list=WRS_2_Path_Row_list,
        WRS_1_Path_Row_list=WRS_1_Path_Row_list,
        archive_path=ur"S:\Landsat",
        temp_dwnld_folder=ur'G:\temp'
    )

    # LANDSAT_8_C1          Landsat 8 Operational Land Imager and Thermal Infrared Sensor Collection 1 Level-1
    # LANDSAT_8_PREWRS      Landsat 8 Operational Land Imager and Thermal Infrared Sensor Pre-WRS-2: 2013
    # LANDSAT_ETM_C1        Landsat 7 Enhanced Thematic Mapper Plus Collection 1 Level-1
    # LANDSAT_TM_C1         Landsat 4-5 Thematic Mapper Collection 1 Level-1
    # LANDSAT_MSS           Landsat 1-5 Multispectral Scanner: 1972-2013
Camarada Che
fonte
1

Você pode usar o selenium webdriver para acessar e navegar em sites. Você também pode usar o BeatifulSoup para raspar e identificar downloads. Usar esses pacotes juntos resolverá seu problema.

https://www.seleniumhq.org/

https://www.crummy.com/software/BeautifulSoup/bs4/doc/

Aqui está um exemplo de código funcional usando esses pacotes para automatizar o download de dados da web. Você deve poder alterar este script para atender às suas necessidades.

import time
import os
import urllib2
import re
from bs4 import BeautifulSoup
from selenium import webdriver
from selenium.webdriver.support.ui import Select
import requests

#Set Output
output = (r"C:\DataDownloads")

#Get current date
current_date = (time.strftime("%m%d%Y"))


# This script requires a web driver to run, and must be downloaded prior to executing the script
# For instance, chrome webdriver (https://sites.google.com/a/chromium.org/chromedriver/downloads), or webdriver for browser of choice.
# This scripts webdriver is currently set to the chrome browser. 


#                                      **Make sure the webdriver in your PATH or else the script will fail.**
driver = webdriver.Chrome()


##################################################################################################################
Automate website data download
##################################################################################################################

#Web address for the data site
url = "https://data-mrgis.opendata.arcgis.com/datasets/madison-county-parcels-live"

#initialize webdriver
driver.get(url)

# wait till the web page is fully loaded
time.sleep(8)

# make the dropdown options available for scraping
element = driver.find_element_by_id('download-button')
element.click()

# scrape the page in its current state and close browser
content = driver.page_source.encode('utf-8').strip()
driver.close()

# Use BeaurifulSoup to scrape the page data
soup = BeautifulSoup(content,"html.parser")

# Find everything with "li" (list) tags
li_tags = soup.find_all("li")

# create empty list
zip_tags = []

# get a list of li tags that have a zip file path in them
for n in li_tags:
    s = str(n.contents[0])

    if ".zip" in s:
        zip_tags.append(s)

#Find our download link using regex
for n in zip_tags:
    result = re.search('href="(.*)" id', n)
    dwnld_url = result.group(1)

#Check if the output directory exists. If not, than create new directory.
Idaho = os.path.join(output, "ID")
if not os.path.exists(Idaho):
    os.makedirs(Idaho)

#Create path for data export
Complete_Path = os.path.join(Idaho, "Madison_" + current_date + ".zip")

#Read and Write download to file output location
with open(Complete_Path, "wb") as Madison:
    ID_data = urllib2.urlopen(dwnld_url)
    ID_data_write = ID_data.read()
    print ("Downloading Data")
Pdavis327
fonte
0

Eu uso o aplicativo BDA para baixar cenas inteiras. Contudo; Recentemente, tenho usado a API python do Google Earth Engine e um módulo em https://github.com/loicdtx/landsat-extract-gee e funciona muito bem para a extração de um único pixel, mas é fácil fazer cenas inteiras. Fácil de configurar, documentado com decência, basta seguir o protocolo de aceitação da "lista de permissões" ou você receberá o Erro 403.

planetsandman
fonte