R： 将光栅聚合到shapefile多边形

Question

我想将光栅数据聚合到自定义形状文件中的每个多边形。

在这种情况下，我想获得撒哈拉以南非洲次国家区域城市化的平均程度。

我的sf如下所示：

> africa_map
Simple feature collection with 543 features and 4 fields
Geometry type: MULTIPOLYGON
Dimension:     XY
Bounding box:  xmin: -25.36042 ymin: -46.96575 xmax: 63.49391 ymax: 27.66147
Geodetic CRS:  WGS 84
First 10 features:
    cname ccode        regname continent                       geometry
1  Angola    AO          Bengo    Africa MULTIPOLYGON (((13.371 -8.5...
2  Angola    AO       Benguela    Africa MULTIPOLYGON (((12.53336 -1...
3  Angola    AO            Bie    Africa MULTIPOLYGON (((16.61158 -1...
4  Angola    AO        Cabinda    Africa MULTIPOLYGON (((12.78266 -4...
5  Angola    AO Cuando Cubango    Africa MULTIPOLYGON (((21.9838 -16...
6  Angola    AO   Cuanza Norte    Africa MULTIPOLYGON (((15.40788 -7...
7  Angola    AO     Cuanza Sul    Africa MULTIPOLYGON (((13.7926 -11...

或绘制：

另一方面，光栅数据采用以下形式：

> imported_raster
class      : RasterLayer 
dimensions : 18000, 36082, 649476000  (nrow, ncol, ncell)
resolution : 1000, 1000  (x, y)
extent     : -18041000, 18041000, -9e+06, 9e+06  (xmin, xmax, ymin, ymax)
crs        : +proj=moll +lon_0=0 +x_0=0 +y_0=0 +datum=WGS84 +units=m +no_defs 
names      : GHS_BUILT_LDS1975_GLOBE_R2018A_54009_1K_V2_0 
values     : 0, 100  (min, max)

这些比整个星球所需的要细得多。为了加速计算，我首先聚合光栅，然后将其转换为shapefile，剩余的每个光栅像素都转换为shapefile中的点几何形状。然后，这个shapefile可以聚合到我的区域边界。诚然，这不是很优雅（最终也不起作用）。

library(tidyverse)
library(sf)
library(raster)
library(stars)
library(rgdal)

> # aggregate (to 25x25 km)
> imported_raster_ag <- aggregate(imported_raster, fact=25)
> 
> # convert to sp
> urbanized = rasterToPolygons(imported_raster_ag) 
> # convert to sf
> urbanized_sf <- st_as_sf(urbanized)

# compare projection
st_crs(africa_map)==st_crs(urbanized_sf)
# align projection
urbanized_sf <- st_transform(urbanized_sf, st_crs(africa_map))

urbanized_sf <- urbanized_sf %>% rename(urbanization = GHS_BUILT_LDS1975_GLOBE_R2018A_54009_1K_V2_0)

> urbanized_sf
Simple feature collection with 398872 features and 1 field (with 500 geometries empty)
Geometry type: POLYGON
Dimension:     XY
Bounding box:  xmin: -179.9999 ymin: -57.40086 xmax: 179.9963 ymax: 82.33738
Geodetic CRS:  WGS 84
First 10 features:
   urbanization                       geometry
1             0 POLYGON ((-117.1367 82.3373...
2             0 POLYGON ((-116.2261 82.3373...
3             0 POLYGON ((-115.3156 82.3373...
4             0 POLYGON ((-114.405 82.33738...
5             0 POLYGON ((-113.4944 82.3373...
6             0 POLYGON ((-112.5838 82.3373...
7             0 POLYGON ((-111.6732 82.3373...
8             0 POLYGON ((-110.7627 82.3373...
9             0 POLYGON ((-109.8521 82.3373...
10            0 POLYGON ((-108.9415 82.3373...

当时的想法是，到目前为止，我可以沿着其他旧金山的区域边界将这些点聚合起来。但是，我收到一条错误消息，我找到的唯一建议修复程序只会重复它。

> urbanized_africa <- aggregate(urbanized_sf["urbanization"], by = africa_map$geometry, mean)
although coordinates are longitude/latitude, st_intersects assumes that they are planar
Error in CPL_geos_binop(st_geometry(x), st_geometry(y), op, par, pattern,  : 
  Evaluation error: IllegalArgumentException: Invalid number of points in LinearRing found 2 - must be 0 or >= 4.

> urbanized_sf_fixed <- sf::st_make_valid(urbanized_sf)
Error in CPL_geos_make_valid(x) : 
  Evaluation error: IllegalArgumentException: Invalid number of points in LinearRing found 2 - must be 0 or >= 4.

我想在我的转化过程中，可能有什么地方被破坏了，或者有其他更根本的缺陷。将光栅数据聚合到shapefile多边形中，有什么更优雅、更具功能性的工作流？

Answer 1

查看提取功能

在你的情况下

africamap$urbanized <- extract(imported_raster, africamap, fun="mean")

africa amap仍然应该首先投影到光栅投影。如果您有nodata值，那么明智的做法是向调用添加na.rm=TRUE参数。

为了缩短处理时间，您还可以将光栅裁剪到非洲。

在exactextractr软件包中还有一个更快的extract版本。您还可以利用使用terra而不是光栅来进行光栅处理。