RotationDetect {ICvectorfields}

Detect Rotating Patterns in Vector Fields

Description

Detect patterns in vector fields represented on a grid by looking in the rook's neighbourhood of each grid cell. This function is analogous to PatternDetect, except that it detects rotational patterns. Four patterns are detected: clockwise rotation when rotation in all four neighbbour grids appears clockwise, counter clockwise rotation when rotation in all four neighbour grids appears counter-clockwise, and partial clockwise and counter-clockwise rotation, when all but one of the four adjacent neighbour cells has vectors that indicate rotation. For all of the patterns above a sub-pattern is specified as convergence when all of the vectors in the four adjacent grids point towards the focal cell or a divergence when all of the vectors in the four adjacent grids point away from the focal cell.

Usage

RotationDetect(vfdf)

Arguments

vfdf

A data frame as returned by DispField, DispFieldST, or DispFieldSTall with at least five rows (more is better)

Value

A data frame as returned by DispField, DispFieldST, or DispFieldSTall, with three additional columns. The first additional column is called Pattern in which the patterns around each focal cell are categorized as clockwise, counter-clockwise, partial clockwise, partial counter-clockwise, or NA. The second additional column, called SubPattern, indicates whether all arrows point towards (convergence) or away (divergence) from the focal cell. The third additional column is called PatternCt, which contains a one if all four neighbourhood grid cells contain displacement estimates, and a NA otherwise.

Examples

# creating rotation patterns
Mat1 <- matrix(rep(0,9*9), nrow = 9)
Mat1[c(1:3), 4] <- 1
Mat1[c(7:9), 6] <- 1
Mat1[4, c(7:9)] <- 1
Mat1[6, c(1:3)] <- 1
Mat1

Mat2 <- matrix(rep(0,9*9), nrow = 9)
Mat2[c(1:3), 5] <- 1
Mat2[c(7:9), 5] <- 1
Mat2[5, c(7:9)] <- 1
Mat2[5, c(1:3)] <- 1
Mat2

# rasterizing
rast1 <- terra::rast(Mat1)
terra::plot(rast1)
rast2 <- terra::rast(Mat2)
terra::plot(rast2)

# Detecting clockwise rotation
(VFdf1 <- DispField(rast1, rast2, factv1 = 3, facth1 = 3, restricted = TRUE))
(patdf1 <- RotationDetect(VFdf1))

# Detecting counter-clockwise rotation
(VFdf2 <- DispField(rast2, rast1, factv1 = 3, facth1 = 3, restricted = TRUE))
(patdf2 <- RotationDetect(VFdf2))