python rasterio 基于矢量裁剪栅格包括属性筛选

之前搜索了rasterio的矢量裁剪的方法，但是真正运用的时候报错ValueError: No valid geometry objects found for rasterize，但是实际上我的shp和raster是重叠的，发现是没有将矢量数据的投影与栅格的进行统一。因此采用下面的方法即可：

import fiona
import rasterio
from rasterio.mask import mask
from geopandas import GeoDataFrame
from rasterio import features
from shapely.geometry import shape
from scipy import ndimage


try:
    import gdal
except:
    from osgeo import gdal
    
def extract_by_shp_rasterio(in_shp_path, in_raster_path, out_raster_path):
    maskdata = GeoDataFrame.from_file(in_shp_path)
    
    # crop即保留shp里的,invert=True即去除shp内的
    with rasterio.open(in_raster_path) as src:
        shpdata = maskdata.to_crs(src.crs)  # 投影转换，使矢量数据与栅格数据投影参数一致
        geoms = [feature for feature in shpdata.geometry]
        out_image, out_transform = mask(src, geoms, crop=True, nodata=4,all_touched=True)
        out_meta = src.meta.copy()
  
    # 更新输出文件的元数据
    out_meta.update({"driver": "GTiff",
                     "height": out_image.shape[1],
                     "width": out_image.shape[2],
                     "transform": out_transform})
    
    # 保存
    with rasterio.open(out_raster_path, "w", **out_meta) as dest:
        dest.write(out_image)    

另外的一个需求是需要筛选shp的属性，比如说本次我需要FP这个属性=='80'，因此代码做了改动，注意直接shpdata.属性名即可获取属性

def extract_by_landfastshp_rasterio(in_shp_path, in_raster_path, out_raster_path):
    maskdata = GeoDataFrame.from_file(in_shp_path)
    # with fiona.open(in_shp_path, "r", encoding='utf-8') as shapefile:
    #     # 获取所有要素feature的形状geometry
    #     geoms = [feature["geometry"] for feature in shapefile]
    
    
    # 裁剪
    with rasterio.open(in_raster_path) as src:
        shpdata = maskdata.to_crs(src.crs)  # 投影转换，使矢量数据与栅格数据投影参数一致
        # geoms = [feature for feature in shpdata if feature.FP=='80']
        geoms=[]
        for i in range(0, len(shpdata) ):
            geo = shpdata.geometry[i] #获取空间属性，即GeoSeries
            name = shpdata.FP[i]      #获得属性数据，即字段
            if name =='08':
                geoms.append(geo)
        # geoms = [feature for feature in shpdata.geometry if feature.FP=='80']
        out_image, out_transform = mask(src, geoms, invert=True, nodata=6,all_touched=True)
        out_meta = src.meta.copy()
    
    
    
    
    # 更新输出文件的元数据
    out_meta.update({"driver": "GTiff",
                     "height": out_image.shape[1],
                     "width": out_image.shape[2],
                     "transform": out_transform})
    
    # 保存
    with rasterio.open(out_raster_path, "w", **out_meta) as dest:
        dest.write(out_image)