CleanRfield
cleanRfield: This package is a compilation of functions to clean and filter observations from yield monitors or other agricultural spatial point data. Yield monitors are prone to error, and filtering the observations or removing observations from near field boundaries can improve estimates of whole-field yield, combine speed, grain moisture, or other parameters. In this package, users can easily select filters thresholding for one or more traits and prepare a smaller dataset to make decisions.
Install / Use
/learn @filipematias23/CleanRfieldREADME
cleanRfield: A tool for cleaning and filtering data spatial points from crop yield maps in R.
This package is a compilation of functions to clean and filter observations from yield monitors or other agricultural spatial point data. Yield monitors are prone to error, and filtering the observations or removing observations from near field boundaries can improve estimates of whole-field yield, combine speed, grain moisture, or other parameters of interest. In this package, users can easily select filters for one or more traits and prepare a smaller dataset to make decisions.
<p align="center"> <img src="https://raw.githubusercontent.com/filipematias23/images/master/readme/CleanRField_large.jpg" width="40%" height="40%"> </p> <div id="menu" />This tutorial assumes that readers have a basic understanding of spatial data, including projections and coordinate reference systems. If you need a refresher on this topic, we recommend reading this blog post for deciding between projected and unprojected data and this post for understanding the basics of coordinate reference systems.
Resources
2. Cropping or selecting regions
6. Buffering the field boundaries
7. Filtering using data values
8. Filtering using standard deviation values
9. Evaluating multiple fields on parallel
12. Working with .csv or .txt files
<div id="instal" /><br />
Now install R/cleanRfield using the
install_github()function from devtools package. If necessary, use the argument type="source".
install.packages("devtools")
devtools::install_github("filipematias23/cleanRfield")
1. First steps
Calling necessary packages:
library(terra)
library(cleanRfield)
Start this tutorial by downloading the example EX1 here and the field boundary here. This tutorial will use the function
vect()from package terra to read and upload the data to RStudio (see provided code).
EX1 is a yield map from a soybean field, stored as a point shapefile. Yield monitor observations were originally collected in the north-central US using a combine yield monitor, and observations were geographically shifted to protect the landowner’s privacy. This data set include three attributes:
- Speed (miles per hour): speed of the combine at the time the observation was recorded
- Dry_Yield (bushels per acre): yield at that observation’s location as recorded by the yield monitor, adjusted to a 13% moisture basis
- Adj_Moist (percent): indicates what moisture level the original wet yield measurements were adjusted to when calculating dry yield
The field boundary is a shapefile layer with three (3) polygons that delineate the boundaries of the sample field EX1.
### Opening Sample Field 1 ###
par(mfrow=c(1,2))
EX1<-vect("EX1.shp")
plot(EX1, main="Data Point")
EX1.Shape<-vect("EX1_boundary.shp")
plot(EX1.Shape, main="Field Boundary")
par(mfrow=c(1,1))
2. Cropping or selecting targeted field
Users can subset the data by drawing boundaries around a field or subset of fields. Function
cropField. Depending on your computer and the size of your data set, this step may take a few seconds. This function works better using a popup window, not RStudio's integrated viewing pane, so we've included code for opening that new window.
#Open an extra plot window
x11()
# "Use cursor to select 4 points around polygon (1) in the plots window."
EX1.C<-cropField(field = EX1, nPolygon = 1, nPoint = 4)
plot(EX1.C$shape,main="Drawing Shape")
# Using the shape drawn above to crop data:
EX1.C<-cropField(field = EX1, shape = EX1.C$shape)
# All data will be selected using the full boundary as shape:
EX1.C1<-cropField(field = EX1, shape = EX1.Shape)
The newest version of RStudio (2021.09.2) has updated the plot viewing pane. If you are using the newest RStudio, you may need run an additional line of code to open the point-and-click cropping functionality in a pop-out window instead of the integrated plot viewing pane.
#Open an extra plot window
x11()
# "Use cursor to select 4 points around polygon (1) in the plots window."
EX1.C<-cropField(field = EX1, nPolygon = 1, nPoint = 4)
3. Sampling point data
Users can sample random points in the data. Function
sampleField.
# Sampling 5%:
EX1.S<-sampleField(field = EX1, size = 0.05)
# Sampling 10% under a small shape:
EX1.S<-sampleField(field = EX1,shape = EX1.C$shape, size = 0.1)
# Sampling 10% under a full shape:
EX1.S<-sampleField(field = EX1,shape = EX1.Shape, size = 0.1)
4. Making rasters
Function
rasterField.Data points can be used to create raster files. Use either the provided code for unprojected data or projected data-- you will not need to run both sets of code. Any positive number can be chosen as the resolution, but choosing too high of a resolution will result in a raster file that oversimplifies the shape of the field, and choosing too low of a resolution can cause the runtime to be to long and/or cause the parts of the field between combine passes to be excluded from the final field shape.
# Check projection to observe 'LENGTHUNIT':
crs(EX1)
# Unprojected Data (non or NA): use resolution around 0.00008 to create a raster for "Dry_Yield":
EX1.R<-rasterField(field = EX1,
trait = c("Dry_Yield"),
res = 0.00008)
# Projected Data (e.g., +units=m or +units=us-ft): use resolution around 5 to 20 to create a raster for "Dry_Yield":
EX1.R<-rasterField(field = EX1,
trait = c("Dry_Yield"),
res = 20)
# Making raster only for the small shape:
EX1.R<-rasterField(field = EX1,
shape = EX1.C$shape,
trait = c("Dry_Yield"),
res = 0.00008) # Attention: for projected data use res=20 (e.g., +units=m or +units=us-ft).
# Multilayer raster for two or more traits:
EX1.R<-rasterField(field = EX1,
trait = c("Dry_Yield","Speed"),
res = 0.00008) # Attention: for projected data use res=20 (e.g., +units=m or +units=us-ft).
# Different raster color visualizations:
library(RColorBrewer)
par(mfrow=c(2,3))
plot(EX1.R$Dry_Yield)
plot(EX1.R$Dry_Yield,col = heat.colors(10))
plot(EX1.R$Dry_Yield,col = topo.colors(10))
plot(EX1.R$Dry_Yield,col = brewer.pal(11, "RdYlGn"))
plot(EX1.R$Dry_Yield,col = brewer.pal(9, "BuGn"))
plot(EX1.R$Dry_Yield,col = brewer.pal(9, "Greens"))
par(mfrow=c(1,1))
5. Building shape boundaries
Users can manually draw field boundaries or use the ras
