A Computational Introduction to Digital Image Processing, 2nd Edition

110 A Computational Introduction to Digital Image Processing, Second Edition

(a) High boost filtering with hb1 (b) High boost filtering with hb2

FIGURE 5.18: High boost filtering

and Figure 5.18(b) similarly.

With Python, these images could be obtained easily as:

Python

In: k = im2fl(io.imread(’koala.png’))

In: kf = ndi.uniform

_

filter(k,3)

In: hb1 = 3

*

k - 2

*

kf

In: hb2 = 1.25

*

k - 0.25

*

kf

In: subplot(121),io.imshow(hb1,vmax=1.0,vmin=0.0)

In: subplot(122),io.imshow(hb2,vmax=1.0,vmin=0.0)

Of the two filters, hb1 appears to produce the best result; hb2 produces an image not much

crisper than the original.

5.8 Non-Linear Filters

Linear filters, as we have seen in the ...