AI Earth ——开发者模式案例6：决策树模型实现冬小麦提取

import aie
aie.Authenticate()
aie.Initialize()

feature_collection = aie.FeatureCollection('China_City') \
                        .filter(aie.Filter.eq('city', '亳州市'))
region = feature_collection.geometry()

# 指定检索数据集，可设置检索的空间和时间范围，以及属性过滤条件（如云量过滤等）
def getl8_ndvi(startdate,enddate):
    dataset = aie.ImageCollection('LANDSAT_LC08_C02_T1_L2') \
                 .filterBounds(region) \
                 .filterDate(startdate, enddate) 
                 # .filter(aie.Filter.lte('eo:cloud_cover', 20.0))
    ndvi =  dataset.map(get_ndvi)  
    return ndvi
def get_ndvi(image):
    ndvi = image.normalizedDifference(['SR_B5', 'SR_B4'])
    return ndvi  
# 黄淮海地区冬小麦典型物候期。播种期 10-11月，旺长期3-4月，成熟期5-6月
NDVI_median = getl8_ndvi('2017-10-11', '2017-11-10').median().clip(region) 
NDVI_max = getl8_ndvi('2018-03-20', '2018-04-20').max().clip(region) 
NDVI_min = getl8_ndvi('2018-05-20', '2018-06-30').min().clip(region) 
# 冬小麦提取规则集
mask1 = NDVI_max.gt(aie.Image.constant(0.33))     # 过滤水体、不透水面、裸地等非植被区域
mask2 = NDVI_median.lt(aie.Image.constant(0.50))  # 过滤森林、春播作物
mask3 = NDVI_max.lte(aie.Image.constant(0.48))    # < 0.48存在
mask4 = NDVI_max.gt(aie.Image.constant(0.48))     # 区分植被区域和非植被区域，> 0.48 为植被区域
mask5 = NDVI_min.gt(aie.Image.constant(-0.12))    # 筛选冬小麦区域，成熟/收获期冬小麦植被指数下降
mask6 = NDVI_min.lt(aie.Image.constant(0.17))   
mask7 = NDVI_min.lt(NDVI_max.add(aie.Image.constant(0.5)))
wheat1 = mask1.And(mask2).And(mask3).And(mask5).And(mask6)
wheat2 = mask1.And(mask2).And(mask4).And(mask5).And(mask7)
wheat  = wheat1.add(wheat2).where((wheat1.add(wheat2)).gt(aie.Image.constant(0)),aie.Image.constant(1))

map = aie.Map(
    center=region.getCenter(),
    height=800,
    zoom=7
)
vis_params = {
    'color': '#00FF00'
}
map.addLayer(
    region,
    vis_params,
    'region',
    bounds=region.getBounds()
)
mask_vis  = {
    'min': 0,
    'max': 1,
    'palette': ['#ffffff', '#008000']    # 0:白色, 1:绿色
}
ndvi_vis  = {
    'min': -0.2,
    'max': 0.6,
    'palette': ['#d7191c', '#fdae61', '#ffffc0', '#a6d96a', '#1a9641']
}
map.addLayer(NDVI_median,ndvi_vis, 'NDVI_median', bounds=region.getBounds())
map.addLayer(NDVI_max,ndvi_vis, 'NDVI_max', bounds=region.getBounds())
map.addLayer(NDVI_min,ndvi_vis, 'NDVI_min', bounds=region.getBounds())
map.addLayer(wheat,mask_vis, 'wheat', bounds=region.getBounds())    # 绿色区域为小麦
map