Skip to contents

Medial Axis Transform (Internal)

Source: R/Thinning4.R
medial_axis_transform.Rd

This internal function computes the medial axis transform of a binary image, identifying the set of skeleton points equidistant to the object's boundaries.

Usage

medial_axis_transform(img, verbose = FALSE, select.layer = NULL)

Arguments

img

A matrix, data frame, or `SpatRaster` object representing the binary image for transformation.

verbose

Logical. If `TRUE`, outputs diagnostic information such as image dimensions, progress of computation, and final skeleton size. Default is `FALSE`.

select.layer

Integer indicating the layer to use if `img` is a multi-layer `SpatRaster`. Default is 2.

Value

A binary matrix where `1` represents skeleton pixels and `0` represents the background.

Details

- The input image is first processed using load_flexible_image to ensure it is a binary matrix. - The algorithm proceeds through the following steps: 1. **Distance Transform**: Computes the distance of each foreground pixel to the nearest background pixel using a two-pass algorithm. 2. **Local Maxima Detection**: Identifies local maxima in the distance transform to mark potential skeleton points. 3. **Skeleton Refinement**: Ensures connectivity by connecting skeleton points within an 8-neighborhood. - The result is a binary image representing the medial axis of the input object.

Examples

if (FALSE) { # \dontrun{
# Example usage
raster <- terra::rast(matrix(c(0, 1, 1, 0, 0, 1, 1, 0,1), nrow = 3))
skeleton <- medial_axis_transform(raster, verbose = TRUE)
} # }