Test conditions for neighbors and neighbors of neighbors

Evaluate conditions for cells neighboring specific classes and classify them if conditions are true.

cond.4.nofn(
  attTbl,
  ngbList,
  rNumb = FALSE,
  classVector,
  class,
  nbs_of,
  cond,
  min.bord = NULL,
  max.iter = +Inf,
  peval = 1,
  directional = FALSE,
  ovw_class = FALSE,
  hgrowth = FALSE
)

Arguments

attTbl: data.frame, the attribute table returned by the function attTbl.
ngbList: list, the list of neighborhoods returned by the function ngbList.
rNumb: logic, the neighborhoods of the argument ngbList are identified by cell numbers (rNumb=FALSE) or by row numbers (rNumb=TRUE) (see ngbList). It is advised to use row numbers for large rasters.
classVector: numeric vector, defines the cells in the attribute table that have already been classified. See conditions for more information about class vectors.
class: numeric, the classification number to assign to all cells that meet the function conditions.
nbs_of: numeric or numeric vector, indicates the class(es) of focal and anchor cells. Conditions are only evaluated at positions adjacent to anchor and focal cells. If the classification number assigned with the argument class is also included in the argument nbs_of, the function takes into account class continuity (see conditions).
cond: character string, the conditions a cell have to meet to be classified as indicated by the argument class. The classification number is only assigned to unclassified cells unless the argument ovw_class = TRUE. See conditions for more details.
min.bord: numeric value between 0 and 1. A test cell is classified if conditions are true and if among its bordering cells a percentage equal or greater than min.bord belong to one of the classes of nbs_of. Percentages are computed counting only valid neighbors (i.e., neighbors with complete cases).
max.iter: integer, the maximum number of iterations.
peval: numeric value between 0 and 1. If absolute or relative neighborhood conditions are considered, test cells are classified if the number of positive evaluations is equal or greater than the percentage specified by the argument peval (see conditions).
directional: logic, absolute or relative neighborhood conditions are tested using the directional neighborhood (see conditions).
ovw_class: logic, reclassify cells that were already classified and that meet the function conditions.
hgrowth: logic, if true the classes in nbs_of are treated as discrete raster objects and the argument class is ignored.

Value

Update classVector with the new cells that were classified by the function. See conditions for more details about class vectors.

Details

The function evaluates the conditions of the argument cond for all unclassified cells in the neighborhood of focal and anchor cells (specified by the argument nbs_of). Unclassified cells are NA-cells in classVector.
Cells that meet the function conditions are classified as indicted by the argument class.
Class continuity is considered if the classification number assigned with the argument class is also included in the argument nbs_of. This means that, at each iteration, newly classified cells become focal cells and conditions are tested in their neighborhood.
All types of conditions can be used. The condition string can only include one neighborhood condition ('{}') (see conditions).

Homogeneous growth (hgrowth)

If the argument hgrowth is true the classes in nbs_of are treated as discrete raster objects and the argument class is ignored. Iterations proceed as follow:

cells contiguous to the first element of nbs_of are evaluated against the classification rules and, when evaluations are true, cells are assigned to that element;
the same process is repeated for cells contiguous to the second element of nbs_of, then for cells contiguous to the third element and so on until the last element of nbs_of;
once cells contiguous to the last element of nbs_of are evaluated the iteration is complete;
cells classified in one iteration become focal cells in the next iteration;
a new iteration starts as long as new cells were classified in the previous iteration and if the iteration number < max.iter.

Examples

# DUMMY DATA
######################################################################################
# LOAD LIBRARIES
library(scapesClassification)
library(terra)

# LOAD THE DUMMY RASTER
r <- list.files(system.file("extdata", package = "scapesClassification"),
                pattern = "dummy_raster\\.tif", full.names = TRUE)
r <- terra::rast(r)

# COMPUTE THE ATTRIBUTE TABLE
at <- attTbl(r, "dummy_var")

# COMPUTE THE LIST OF NEIGBORHOODS
nbs <- ngbList(r)

# SET A DUMMY FOCAL CELL (CELL #25)
at$cv[at$Cell == 25] <- 0

# SET FIGURE MARGINS
m <- c(2, 8, 2.5, 8)

######################################################################################
# ABSOLUTE TEST CELL CONDITION - NO CLASS CONTINUITY
######################################################################################

# conditions: "dummy_var >= 3"
cv1 <- cond.4.nofn(attTbl = at, ngbList = nbs,

                   # CLASS VECTOR - INPUT
                   classVector = at$cv,

                   # CLASSIFICATION NUMBER
                   class = 1,

                   # FOCAL CELL CLASS
                   nbs_of = 0,

                   # ABSOLUTE TEST CELL CONDITION
                   cond = "dummy_var >= 3")

# CONVERT THE CLASS VECTOR INTO A RASTER
r_cv1 <- cv.2.rast(r, at$Cell,classVector = cv1, plot = FALSE)

# PLOT
plot(r_cv1, type="classes", axes=FALSE, legend = FALSE, asp = NA, mar = m,
     colNA="#818792", col=c("#78b2c4", "#cfad89"))
text(r)
mtext(side=3, line=1, adj=0, cex=1, font=2, "CONDITION: ABSOLUTE TEST CELL")
mtext(side=3, line=0, adj=0, cex=1, "Class continuity: NO")
mtext(side=1, line=0, cex=0.9, adj=0, "Rule: 'dummy_var >= 3'")
legend("bottomright", bg = "white", fill = c("#78b2c4", "#cfad89", "#818792"),
       legend = c("Focal cell", "Classified cells", "Unclassified cells"))


######################################################################################
# ABSOLUTE TEST CELL CONDITION - WITH CLASS CONTINUITY
######################################################################################

# conditions: "dummy_var >= 3"
cv2 <- cond.4.nofn(attTbl = at, ngbList = nbs, classVector = at$cv,

                  # CLASSIFICATION NUMBER
                   class = 1,

                   nbs_of = c(0,  # FOCAL CELL CLASS
                              1), # CLASSIFICATION NUMBER

                   # ABSOLUTE CONDITION
                   cond = "dummy_var >= 3")

# CONVERT THE CLASS VECTOR INTO A RASTER
r_cv2 <- cv.2.rast(r, at$Cell,classVector = cv2, plot = FALSE)

# PLOT
plot(r_cv2, type="classes", axes=FALSE, legend = FALSE, asp = NA, mar = m,
     colNA="#818792", col=c("#78b2c4", "#cfad89"))
text(r)
mtext(side=3, line=1, adj=0, cex=1, font=2, "CONDITION: ABSOLUTE TEST CELL")
mtext(side=3, line=0, adj=0, cex=1, "Class continuity: YES")
mtext(side=1, line=0, cex=0.9, adj=0, "Rule: 'dummy_var >= 3'")
legend("bottomright", bg = "white", fill = c("#78b2c4", "#cfad89", "#818792"),
       legend = c("Focal cell", "Classified cells", "Unclassified cells"))


######################################################################################
# ABSOLUTE NEIGHBORHOOD CONDITION
######################################################################################

# conditions: "dummy_var{} >= 3"
cv3 <- cond.4.nofn(attTbl = at, ngbList = nbs, classVector = at$cv, nbs_of = c(0,1), class = 1,

                   # ABSOLUTE NEIGHBORHOOD CONDITION
                   cond = "dummy_var{} >= 3",

                   # RULE HAS TO BE TRUE FOR 100% OF THE EVALUATIONS
                   peval = 1)

# CONVERT THE CLASS VECTOR INTO A RASTER
r_cv3 <- cv.2.rast(r, at$Cell,classVector = cv3, plot = FALSE)

#PLOT
plot(r_cv3, type="classes", axes=FALSE, legend = FALSE, asp = NA, mar = m,
     colNA="#818792", col=c("#78b2c4", "#cfad89"))
text(r)
mtext(side=3, line=1, adj=0, cex=1, font=2, "CONDITION: ABSOLUTE NEIGHBORHOOD")
mtext(side=3, line=0, adj=0, cex=1, "Class continuity: YES")
mtext(side=1, line=0, cex=0.9, adj=0, "Rule: 'dummy_var{ } >= 3'")
mtext(side=1, line=0, cex=0.9, adj=1, "('{ }' cell neighborhood)")
mtext(side=1, line=1, cex=0.9, adj=0, "Fn_perc: 1 (100%)")
legend("bottomright", bg = "white", fill = c("#78b2c4", "#cfad89", "#818792"),
       legend = c("Focal cell", "Classified cells", "Unclassified cells"))


######################################################################################
# RELATIVE NEIGHBORHOOD CONDITION
######################################################################################

# conditions: "dummy_var > dummy_var{}"
cv4 <- cond.4.nofn(attTbl = at, ngbList = nbs, classVector = at$cv, nbs_of = c(0,1), class = 1,

                   # RELATIVE NEIGHBORHOOD CONDITION
                   cond = "dummy_var > dummy_var{}",

                   # RULE HAS TO BE TRUE FOR AT LEAST 60% OF THE EVALUATIONS
                   peval = 0.6)


# CONVERT THE CLASS VECTOR INTO A RASTER
r_cv4 <- cv.2.rast(r, at$Cell, classVector = cv4, plot = FALSE)

#PLOT
plot(r_cv4, type="classes", axes=FALSE, legend = FALSE, asp = NA, mar = m,
     colNA="#818792", col=c("#78b2c4", "#cfad89"))
text(r)
mtext(side=3, line=1, adj=0, cex=1, font=2, "CONDITION: RELATIVE NEIGHBORHOOD")
mtext(side=3, line=0, adj=0, cex=1, "Class continuity: YES")
mtext(side=1, line=0, cex=0.9, adj=0, "Rule: 'dummy_var > dummy_var{ }'")
mtext(side=1, line=0, cex=0.9, adj=1, "('{ }' cell neighborhood)")
mtext(side=1, line=1, cex=0.9, adj=0, "Fn_perc: 0.6 (60%)")
legend("bottomright", bg = "white", fill = c("#78b2c4", "#cfad89", "#818792"),
       legend = c("Focal cell", "Classified cells", "Unclassified cells"))


######################################################################################
# RELATIVE FOCAL CELL CONDITION
######################################################################################

# conditions: "dummy_var > dummy_var[]"
cv5 <- cond.4.nofn(attTbl = at, ngbList = nbs, classVector = at$cv, nbs_of = c(0,1), class = 1,

                   # RELATIVE FOCAL CELL CONDITION
                   cond = "dummy_var > dummy_var[]")


# CONVERT THE CLASS VECTOR INTO A RASTER
r_cv5 <- cv.2.rast(r, at$Cell,classVector = cv5, plot = FALSE)

#PLOT
plot(r_cv5, type="classes", axes=FALSE, legend = FALSE, asp = NA, mar = m,
     colNA="#818792", col=c("#78b2c4", "#cfad89"))
text(r)
mtext(side=3, line=1, adj=0, cex=1, font=2, "CONDITION: RELATIVE FOCAL CELL")
mtext(side=3, line=0, adj=0, cex=1, "Class continuity: YES")
mtext(side=1, line=0, cex=0.9, adj=0, "Rule: 'dummy_var > dummy_var[ ]'")
mtext(side=1, line=0, cex=0.9, adj=1, "('[ ]' focal cell)")
legend("bottomright", bg = "white", fill = c("#78b2c4", "#cfad89", "#818792"),
       legend = c("Focal cell", "Classified cells", "Unclassified cells"))


######################################################################################
# HOMOGENEOUS GROWTH
######################################################################################

# Dummy raster objects 1 and 2
ro <- as.numeric(rep(NA, NROW(at)))
ro[which(at$dummy_var == 10)] <- 1
ro[which(at$dummy_var == 8)] <- 2

# Not homogeneous growth
nhg <- cond.4.nofn(attTbl = at, ngbList = nbs, classVector = ro,
                   nbs_of = 1, class = 1, # GROWTH ROBJ 1
                   cond = "dummy_var <= dummy_var[] & dummy_var != 1")

nhg <- cond.4.nofn(attTbl = at, ngbList = nbs, classVector = nhg, # UPDATE nhg
                   nbs_of = 2, class = 2, # GROWTH ROBJ 2
                   cond = "dummy_var <= dummy_var[] & dummy_var != 1")


# Homogeneous growth
hg <- cond.4.nofn(attTbl = at, ngbList = nbs, classVector = ro,
                  nbs_of = c(1, 2), class = NULL,
                  cond = "dummy_var <= dummy_var[] & dummy_var != 1",
                  hgrowth = TRUE) # HOMOGENEOUS GROWTH

# Convert class vectors into rasters
r_nhg <- cv.2.rast(r, at$Cell,classVector = nhg, plot = FALSE)
r_hg  <- cv.2.rast(r, at$Cell,classVector = hg, plot = FALSE)

# Plots
oldpar <- par(mfrow = c(1,2))
m <- c(3, 1, 5, 1)

# Original raster objects (for plotting)
r_nhg[at$dummy_var == 10] <- 3
r_nhg[at$dummy_var == 8]  <- 4

r_hg[at$dummy_var == 10] <- 3
r_hg[at$dummy_var == 8]  <- 4
#t
# 1)
plot(r_nhg, type="classes", axes=FALSE, legend=FALSE, asp=NA, mar = m,
     colNA="#818792", col=c("#78b2c4", "#cfc1af", "#1088a0", "#cfad89"))
text(r)
mtext(side=3, line=1, adj=0, cex=1, font=2, "RASTER OBJECTS GROWTH")
mtext(side=3, line=0, adj=0, cex=0.9, "Not homogeneous (hgrowth = FALSE)")
mtext(side=1, line=0, cex=0.9, adj=0, "Growth rule:")
mtext(side=1, line=1, cex=0.9, adj=0, "'dummy_var<=dummy_var[ ] & dummy_var!=1''")
legend("topleft", bg = "white", y.intersp= 1.3,
       fill = c("#1088a0", "#cfc1af", "#78b2c4", "#cfc1af", "#818792"),
       legend = c("RO1", "RO2", "RO1 - growth", "RO2 - growth", "Unclassified cells"))
# 2)
plot(r_hg, type="classes", axes=FALSE, legend=FALSE, asp=NA, mar = m,
     colNA="#818792", col=c("#78b2c4", "#cfc1af", "#1088a0", "#cfad89"))
text(r)
mtext(side=3, line=1, adj=0, cex=1, font=2, "RASTER OBJECTS GROWTH")
mtext(side=3, line=0, adj=0, cex=0.9, "Homogeneous (hgrowth = TRUE)")
mtext(side=1, line=0, cex=0.9, adj=0, "Growth rule:")
mtext(side=1, line=1, cex=0.9, adj=0, "'dummy_var<=dummy_var[ ] & dummy_var!=1''")
legend("topleft", bg = "white", y.intersp= 1.3,
       fill = c("#1088a0", "#cfc1af", "#78b2c4", "#cfc1af", "#818792"),
       legend = c("RO1", "RO2", "RO1 - growth", "RO2 - growth", "Unclassified cells"))

par(oldpar)

Arguments

Value

Details

See also

Examples