I want to create a subset of an image with four bands. Therefore I am using the crop function in R.
A<-raster("L8_stacked.tif")
subset<-extent(c(639451, 660104, 5469254, 5489566))
B<-crop(A,subset)
As a result I get a raster with only one band in the .tif file. Do I have to define other options to get a subset image with 4 bands?
As the others already pointed out in the comments, the raster() function returns a (single) RasterLayer object. If you want a multilayer raster object you need to use the stack() or brick() function load the image into R. I.e.:
A <- stack("L8_stacked.tif")
If you then apply your extent with the crop() function, the result should be a raster stack containing all the bands from the original image.
To learn more on the raster package, read this document.
Related
I there a way to put a grid over point cloud data using the LiDR package in R.
I want to place a 30x30m grid over a lidar data that I have and then run some statistics on each 10x10 square within the grid. Does anyone know if this is possible and what function I would use?
You can use grid_metrics() with a RasterLayer as layout. From the manual:
res numeric. The resolution of the output Raster. Can optionally be a RasterLayer. In that case the RasterLayer is used as the layout.
Package landscapemetrics can calculate area of each patch for a given raster file, shape of that patch and so on. I want to have not only tibble-frame with patch metrics calculated, but a new raster where each pixel within specific patch will have a value of the area of that patch, shape indicator and so on. We can do it with function spatialize_lsm() (it produces a Large list nested object with probably RasterObject objects within):
library(landscapemetrics)
plot(podlasie_ccilc) # this raster data is provided with package
podlasie.metrics.area <- spatialize_lsm(podlasie_ccilc, what = 'lsm_p_area') # creates a list
plot(podlasie.metrics.area) # produces an error...
How to get a desirable raster file with patch metrics from that list? I guess it is a question of raster package or something else, since landscapemetrics documentation tells nothing about this step.
I not that this data and new raster do not have resolution of the pixel like in meters (30, 30 for Landsat satellite image, for example). So we cannot plot the new raster produced:
podlasie.metrics.area[[1]]
plot(podlasie.metrics.area[[1]])
So I guess landscapemetrics cannot deal with such rasters, we can even use its function to check a suitability of the prior raster for patch discovering:
check_landscape(podlasie_ccilc)
Upd. I did it for the Landsat dataset with resolution 30, 30 and it produced patch area raster, but again I cannot open/show/save as raster it, because of the same error.
Package maintainer helps to solve a problem (yes, it is just related to the structure of list):
plot(podlasie.metrics.area[[1]]$lsm_p_area)
I am running into a problem with the "writeRaster" function in the raster package in R. I am importing a raster (TIF) that I made in ArcGIS (a distance to feature raster).
My goal was to resample the distance raster to the correct resolution and extent, then "mask" it with the appropriate raster to crop it to the shape I require. When I check the results of the mask with the basic plot function, everything looks great and I can see that each pixel in the new masked raster has a distance value.
However, when I write this raster to a file using the writeRaster function, the resulting raster looks like "swiss cheese" and has missing values for any distance over 35km. After much reading, I cannot find any documentation to suggest that there is a way to modify the maximum value set by writeRaster---or that it should even be setting a max value. I have included my code and the basic plots below. A big thank you to anyone who attempts to help me with this!
#Read in distance to fresh water raster
distFW <- raster("D:/Academia/Arc Data/Grackle/NicaCR_90mlayers/dist_FW.tif")
[plot(distFW)][1]
#Resample this layer to the desired resolution and template
NiCR_DistFW<-as.integer(resample(distFW,NiCRrast.tmpl,method="ngb"))
#essentially the same as the first plot
[plot(NiCR_DistFW)][2]
#Mask the resampled raster to the desired shape
NiCR.DistFW.mask.utm <- mask(NiCR_DistFW,NiCR_Mask) #with CA countries cut out.
[plot(NiCR.DistFW.mask.utm)][3]
#write raster to file (this is where things get weird)
writeRaster(x=NiCR.DistFW.mask.utm, filename='DistFWmask2.tif', format='GTiff', datatype='INT2S') #a way to ensure INT2S
#read the newly written raster file in to R so we can review it
dFW <-raster("DistFWMask2.tif")
[plot(dFW)_writeRaster_result][4]
[1]: https://i.stack.imgur.com/v9RkK.jpg
[2]: https://i.stack.imgur.com/v2DG3.jpg
[3]: https://i.stack.imgur.com/cCwJe.jpg
[4]: https://i.stack.imgur.com/MjWj7.jpg
As you can see from plot 4, an undesirable max value has been set. I was the raster I write to file to look like the one in plot 3, not plot 4.
Thanks in advance for any advice.
Well friends, after taking an hour to detail my question I managed to figure out the answer myself. It had to do with setting the datatype.
INT2S has a maximum value of 32,767
by switching it to INT4S, I capture the full range of values in my raster.
Problem solved!
I want to clip a raster using shapefile in r as a shape of shapefile. I used crop function but it clipping as extents not as a shape. Please help me
Crop, as stated in the manual pages, crops to the extent of the object,
for your purposes you may use raster::mask(), with the same syntaxis of crop
mask function "Creates a new Raster* object that has the same values as x, except for the cells that are NA (or other maskvalue) in a ’mask’."
You may pass inverse=T so you get those values out of the spatialPolygon.
I was working with spatial data to get ready for analyses - I have a DEM at the desired extent of my study area, though I have ~39 other layers at the national scale (US). Is there a way to crop all of these 39 layers to the same extent as the DEM at once?
Also, I will be overlaying the output with other layers in a different projection. Is it possible to adjust the projection and pixel size of the output layers?
I am trying to use freeware as much as possible for my data manipulation...
I had the problem above, but have written a function in R to do all of this in a batch - see below. I had 39 climate data layers at the scale of the continental U.S. (from PRISM Climate Data group; http://www.prism.oregonstate.edu/), and wanted to clip them to the extent of a DEM, in southern California, reproject them, and export them for easy import and use with other layers in SAGA GIS. Below is the code, with an example of how it would be run, setting the working directory to the folder that has the layers that you want to crop, and only those layers.
During the processing, all data are stored in memory, so with huge datasets, it might get hung up because of lack of memory... that would probably be something that would be good to improve.
Also, a response on the R Forum provided a shorter, more elegant way to do it too: http://permalink.gmane.org/gmane.comp.lang.r.geo/18320
I hope somebody finds it useful!
#########################################
#BatchCrop Function ###
#by Mike Treglia, mtreglia#gmail.com ###
###Tested in R Version 3.0.0 (64-bit), using 'raster' version 2.1-48 and 'rgdal' version 0.8-10
########################################
#This function crops .asc raster files in working directory to extent of another layer (referred to here as 'reference' layer), converts to desired projection, and saves as new .asc files in the working directory. It is important that the original raster files and the reference layer are all in the same projection, though different pixel sizes are OK. The function can easily be modified to use other raster formats as well
#Note, Requires package 'raster'
#Function Arguments:
#'Reference' refers to name of the layer with the desired extent; 'OutName' represents the intended prefix for output files; 'OutPrj' represents the desired output projection; and 'OutRes' represents the desired Output Resolution
BatchCrop<-function(Reference,OutName,OutPrj,OutRes){
filenames <- list.files(pattern="*.asc", full.names=TRUE) #Extract list of file names from working directory
library(raster) #Calls 'raster' library
#Function 'f1' imports data listed in 'filenames' and assigns projection
f1<-function(x,z) {
y <- raster(x)
projection(y) <- CRS(z)
return(y)
}
import <- lapply(filenames,f1,projection(Reference))
cropped <- lapply(import,crop,Reference) #Crop imported layers to reference layer, argument 'x'
#Function 'f2' changes projectection of cropped layers
f2<-function(x,y) {
x<-projectRaster(x, crs=OutPrj, res=OutRes)
return(x)
}
output <- lapply(cropped,f2,OutPrj)
#Use a 'for' loop to iterate writeRaster function for all cropped layers
for(i in (1:max(length(filenames)))){ #
writeRaster(output[[i]],paste(deparse(substitute(OutName)), i), format='ascii')
}
}
#############################################
###Example Code using function 'BatchCrop'###
#############################################
#Data layers to be cropped downloaded from: http://www.prism.oregonstate.edu/products/matrix.phtml?vartype=tmax&view=maps [testing was done using 1981-2010 monthly and annual normals; can use any .asc layer within the bounds of the PRISM data, with projection as lat/long and GRS80]
#Set Working Directory where data to be cropped are stored
setwd("D:/GIS/PRISM/1981-2010/TMin")
#Import Reference Layer
reference<-raster("D:/GIS/California/Hab Suitability Files/10m_DEM/10m DEM asc/DEM_10m.asc")
#Set Projection for Reference Layer
projection(reference) <- CRS("+proj=longlat +ellps=GRS80")
#Run Function [desired projection is UTM, zone 11, WGS84; desired output resolution is 800m]
BatchCrop(Reference=reference,OutName=TMinCrop,OutPrj="+proj=utm +zone=11 +datum=WGS84",OutRes=800)