FarmMapsApiClient_KB34_MAST/FarmmapsBulkSatDownload/ShowGeotiff.r

153 lines
7.2 KiB
R
Raw Normal View History

2021-05-18 16:19:28 +00:00
# ShowGeotiff.r
# I downloaded and calculated the stats for the polygon defined in C:\git\FarmMapsApiClient_WURtest\FarmmapsDataDownload\DataDownloadInput.json
# in which I set "SatelliteBand": "wdvi" and in which in the console I requested image 8 for date '2020-06-14'
library(raster)
library(sf)
library(rgdal)
setwd("C:/workdir/groenmonitor/")
# FarmmapsDataDownload generates two files:
fileGeotiff <- "wenr.tif"
# fileJpg <- "thumbnail.jpg"
# FarmmapsBulkSatDownload generates many files. Here is what I tried when inputing the same field for a number of years using BulkSatDownloadInput.json
# fileGeotiff <- "wheat_fld5641_20210224.tif" # 2 files for year 2021. This file is a nice example having in upperleft corner no data
# fileGeotiff <- "wheat_fld5641_20210331.tif" # 2 files for year 2021
# fileGeotiff <- "wheat_fld5641_20200321.tif" # 14 files for year 2020, earliest
# fileGeotiff <- "wheat_fld5641_20200922.tif" # 14 files for year 2020, latest
# fileGeotiff <- "wheat_fld5641_20190121.tif" # 9 files for year 2019, earliest
# fileGeotiff <- "wheat_fld5641_20191117.tif" # 9 files for year 2019, latest
# 1 file for year 2018, with error message 'End of Central Directory record could not be found' and invalid wheat_fld5641_20180630.zip
# fileGeotiff <- "wheat_fld5641_20170526.tif" # 1 file for year 2017
# Zero files for 2016
lenfilename <- nchar(fileGeotiff)
year <- substr(fileGeotiff,lenfilename-11,lenfilename-8)
imgdate <- substr(fileGeotiff,lenfilename-11,lenfilename-4)
# The thumbnail has the polygon clipped out, has 1 layer, no crs and the mean value is not the mean wdvi we are looking for
# r.thumbnail <- raster(fileJpg)
# plot(r.thumbnail)
# crs(r.thumbnail)
# CRS arguments: NA
# cellStats(r.thumbnail,'mean') #59.91667
stk.wenr <- stack(x=fileGeotiff)
# plot(stk.wenr) shows 5 plots (5 bands)
# I think these are:
# 1. ndvi (since it runs from 0 to 1)
# 2. wdvi (since it runs from 0 to 0.5)
# 3-5: RGB (since they run from 0 to 255)
plot(stk.wenr)
# CRS arguments:
# +proj=sterea +lat_0=52.1561605555556 +lon_0=5.38763888888889 +k=0.9999079 +x_0=155000 +y_0=463000 +ellps=bessel +units=m +no_defs
# Or use st_crs(stk.wenr) to get more info, but no EPSG code in there.
# Likely it is epsg:28992 (Amersfoort)
crs(stk.wenr)
stk.wenr <- projectRaster(stk.wenr, crs = CRS('+init=EPSG:28992'))
crs(stk.wenr)
# Looks the same but strangely, if we don't do projectRaster(stk.wenr, crs = CRS('+init=EPSG:28992')), we find below the bottom left corner of the polygon missing
r.wenr.rd.wdvi <- subset(stk.wenr,2)
dev.off()
plot(r.wenr.rd.wdvi,main=paste("wdvi",imgdate),xlab="RDX",ylab="RDY")
cellStats(r.wenr.rd.wdvi,'mean') #0.1654627
# Same as in downloaded file SatelliteDataStatistics_test_satData_wdvi_2020-06-14.csv
# "mean": 0.16564258790046743
# So FindSatelliteItem in C:\git\FarmMapsApiClient_WURtest\FarmmapsApi\Services\GeneralService.cs returns a a stacked geotiff in
# i.e. the cellstats are calculated from the rectangle
# Furthermore we can see
# shows coordinates in RD
# returns a rectangle, thus the shape of the polygon submitted is not clipped.
# The polygon was provided in WGS84. Let's draw it on top.
# First convert the raster to WGS84
r.wenr.wgs84.wdvi <- projectRaster(r.wenr.rd.wdvi, crs = CRS('+init=EPSG:4326'))
# Draw a polygon on top of the raster
# Example polygon p1
# p1 <- data.frame(id = 1, wkt = 'POLYGON((4.963 52.801, 4.966 52.801, 4.966 52.803, 4.963 52.803, 4.963 52.801))')
# p1 <- st_as_sf(p1, wkt = 'wkt', crs = targetcrs)
# plot(p1,add=TRUE, col="transparent",border="black")
# Draw the polygon on top of the raster
# Polygon p2 from C:\git\FarmMapsApiClient_WURtest\FarmmapsDataDownload\DataDownloadInput.json
p2 <- data.frame(id = 1, wkt = gsub("\n","",'POLYGON((
3.37837807779104 51.3231095796538,
3.38065689232502 51.3212527499355,
3.38022924592256 51.3210683536359,
3.37980548452565 51.3208801127141,
3.37959556105776 51.3207540143696,
3.3793691292654 51.3205959677371,
3.37822219207335 51.3215667913007,
3.37816999925795 51.3216109809456,
3.37646704574705 51.3208025481261,
3.37646695791282 51.3208025061493,
3.37608401443192 51.3206231652693,
3.37607169507628 51.3206173959751,
3.37606021048754 51.320612017601,
3.37582728410659 51.3205029306946,
3.37580409779263 51.3206502985963,
3.37575872019649 51.3207993094705,
3.37575476634361 51.3208122883487,
3.37571181656268 51.3208797459348,
3.3756624532907 51.3209415238446,
3.37557609963811 51.3210110142077,
3.37541089899821 51.3211055871218,
3.37477516102591 51.3214102985009,
3.37473173914127 51.3214311108204,
3.37455904622072 51.3215138815012,
3.37415098054777 51.3217199232877,
3.37313700916272 51.3222422862785,
3.37748824689601 51.3242852920348,
3.37749760805371 51.3242713084009,
3.37811903757028 51.3233437635596,
3.37818758851947 51.3232647797363,
3.37823803668144 51.3232236798646,
3.37837807779104 51.3231095796538))'))
p2.wgs84 <- st_as_sf(p2, wkt = 'wkt', crs = CRS('+init=EPSG:4326'))
# Or other way round, in RD Amersfoort. That looks ok
p2.rd <- st_transform(p2.wgs84, "+init=epsg:28992")
# Have a look at both
# wg84
dev.off()
plot(r.wenr.wgs84.wdvi,main=paste("wdvi",imgdate),xlab="LON",ylab="LAT")
plot(p2.wgs84,add=TRUE, col="transparent",border="red")
# RD
dev.off()
plot(r.wenr.rd.wdvi,main=paste("wdvi",imgdate),xlab="RDX",ylab="RDY")
plot(p2.rd,add=TRUE, col="transparent",border="red")
#Let's clip the polygon
r.wenr.rd.wdvi.pol <- mask(r.wenr.rd.wdvi,p2.rd)
r.wenr.wgs84.wdvi.pol <- mask(r.wenr.wgs84.wdvi,p2.wgs84)
dev.off()
plot(r.wenr.wgs84.wdvi.pol,main=paste("wdvi",imgdate),xlab="LON",ylab="LAT")
plot(p2.wgs84,add=TRUE, col="transparent",border="red")
#That's what we want! Now compare the stats
cellStats(r.wenr.rd.wdvi,'mean') #0.1654627 # Stats from rectangle, RD
cellStats(r.wenr.wgs84.wdvi,'mean') #0.1656399 # Stats from rectangle, WGS84
cellStats(r.wenr.rd.wdvi.pol,'mean') #0.1658702 # Stats from raster clipped by polygon, WGS84
cellStats(r.wenr.wgs84.wdvi.pol,'mean') #0.1658611 # Stats from raster clipped by polygon, WGS84
# Conclusion: in this example little difference, but can be quite different!
# The project FarmmapsNbs can generate a wenr.tif file, application.tif, uptake.tif (in rtest1.uptake.zip)and shape.shp (in rtest1.taskmap.zip)
r.application <- raster("C:/git/FarmMapsApiClient_WURtest/FarmmapsNbs/bin/Debug/netcoreapp3.1/Downloads/application.tif")
dev.off()
plot(r.application)
plot(p2.rd,add=TRUE, col="transparent",border="red")
# The application.tif file is a rectangle (polygon not yet clipped), in projection Amersfoort RD New (EPSG:28992)
r.uptake <- raster("C:/git/FarmMapsApiClient_WURtest/FarmmapsNbs/bin/Debug/netcoreapp3.1/Downloads/uptake.tif")
dev.off()
plot(r.uptake)
plot(p2.rd,add=TRUE, col="transparent",border="red")
# The uptake.tif file is a rectangle (polygon not yet clipped), in projection Amersfoort RD New (EPSG:28992)
shp.wgs84 <- readOGR(dsn="C:/git/FarmMapsApiClient_WURtest/FarmmapsNbs/bin/Debug/netcoreapp3.1/Downloads", layer="shape")
crs(shp.wgs84)
# CRS arguments: +proj=longlat +datum=WGS84 +no_defs
dev.off()
plot(r.wenr.wgs84.wdvi,main="wdvi",xlab="LON",ylab="LAT")
plot(shp.wgs84,add=TRUE, col="transparent",border="black")
plot(p2.wgs84,add=TRUE, col="transparent",border="red")
# The shape file is in WGS84