A Computational Introduction to Digital Image Processing, 2nd Edition

Image Geometry 145

(478, 323)

(129, 283)(129, −283)

(478, −323)

0

FIGURE 6.27: The trapezoid around the building

However, this involves a division, which is in general a slow operation. To rewrite as a

multiplication, we would use replace pixel at (x, y) in the new image would take the place

of a pixel at place

(x, sq(x, y − 352) + 352)

where sq (x ) (“sq” stands for “squash”) is a linear function for which

sq(0) = a/b

sq(r) = 1.

This means that

sq(x) =

b −a

br

(x) +

a

b

.

Here’s how this can be implemented in MATLAB or Octave, assuming that the image has

been read as im:

MATLAB/Octave

>> [r,c] = size(im);

>> z = zeros(r,c);

>> x1 = 129; y1=283; x2=478; y2=323;

>> a