rescale: Rescale coordinates from pixels to real length units

View source: R/cl-handling.R

rescaleR Documentation

Rescale coordinates from pixels to real length units

Description

Most of the time, (x, y) coordinates are recorded in pixels. If we want to have them in mm, cm, etc. we need to convert them and to rescale them. This functions does the job for the two cases: i) either an homogeneous rescaling factor, e.g. if all pictures were taken using the very same magnification or ii) with various magnifications. More in the Details section

Usage

rescale(x, scaling_factor, scale_mapping, magnification_col, ...)

Arguments

x

any Coo object

scaling_factor

numeric an homogeneous scaling factor. If all you (x, y) coordinates have the same scale

scale_mapping

either a data.frame or a path to read such a data.frame. It MUST contain three columns in that order: magnification found in ⁠$fac⁠, column "magnification_col", pixels, real length unit. Column names do not matter but must be specified, as read.table reads with header=TRUE Every different magnification level found in ⁠$fac⁠, column "magnification_col" must have its row.

magnification_col

the name or id of the $fac column to look for magnification levels for every image

...

additional arguments (besides header=TRUE) to pass to read.table if 'scale_mapping' is a path

Details

The i) case above is straightforward, if 1cm is 500pix long on all your pictures, just call rescale(your_Coo, scaling_factor=1/500) and all coordinates will be in cm.

The ii) second case is more subtle. First you need to code in your Coo object, in the fac slot, a column named, say "mag", for magnification. Imagine you have 4 magnifications: 0.5, 1, 2 and 5, we have to indicate for each magnification, how many pixels stands for how many units in the real world.

This information is passed as a data.frame, built externally or in R, that must look like this:

mag   pix    cm
0.5   1304   10
1     921    10
2     816    5
5     1020   5

.

We have to do that because, for optical reasons, the ratio pix/real_unit, is not a linear function of the magnification.

All shapes will be centered to apply (the single or the different) scaling_factor.

Value

a Momocs object of same class

Note

This function is simple but quite complex to detail. Feel free to contact me should you have any problem with it. You can just access its code (type rescale) and reply it yourself.

See Also

Other handling functions: arrange(), at_least(), chop(), combine(), dissolve(), fac_dispatcher(), filter(), mutate(), rename(), rm_harm(), rm_missing(), rm_uncomplete(), rw_fac(), sample_frac(), sample_n(), select(), slice(), subsetize()


vbonhomme/Momocs documentation built on Nov. 13, 2023, 8:54 p.m.