(C) 1997 Serge Segal
Dithering algorithm for image.
In this article we'll discuss
smoothing algorithms for static
optical images. First
We will understand why this is necessary, but
you can then have to disassemble itself
process.
So, why is it necessary?
Everyone who has seen the program "EXTACY MD"
Groups "RUSH", remember the part where
rotates the inscription "XTC". Since it is
rotates rapidly, imperceptibly, that the lines
components of its look, to put it mildly,
not a pretty sight, somehow zubtsevato.
The thing that used a 3D-image editor that does not smooth out
lines. Now all the editors for 3D-graphics on powerful machines
do it. For this purpose a specially modified algorithm
Drawing lines and circles. This is what we
describe in the next issues of our
magazine, and now, about how this problem can be solved without
changing the code editor. You can write your own algorithm
which automatically flatten the line.
The second application of smoothing - correction of enlarged
bitmaps. The problem of angularity Pictures come across anyone
who has worked in the editor "Art Studio" with the increase,
especially with the disproportionate. Such an enlarged picture
should also be smooth.
As all seems to work, this algorithm
In the old tradition we do not consider the complex and
maloprimenimye for "ZX-Spectrum" algorithms. So intelligent
smoothing processes one Zoom in about an hour. The effect, of
course ... But here's the performance like something is pulling
... So I came up with a different algorithm smooth, easy-to
idiocy, but At the same time, working quickly
and giving reasonable effect. Algorithm consists of two parts:
smoothing lines (For vector images) and smoothing the corners
(to enlarge An image). Run both algorithms on the same
principle: scanning the screen in search of
template and replace the template on the more beautiful.
As can be seen, for the first
algorithm, we find
=> Tooth and replace it with
square or corner.
What will change, according to
| | Shows variable DOU
BLE, if 0 - then the tooth
otherwise the square. Recommendation
through each for increased every
mappings square, and for
vector - a nook.
Second pass - smoothing angles works with the matrix 3 for 3.
Similar to the first,
This algorithm is replaced
preserves angles in a triangle
=> Lnye teeth. No
complexity in this
No, the main thing - the right
correctly remove the matrix
and the right to return it
in place. Also had to introduce
forced inversion for some pictures for a while smoothing
(variable INVERT). After the end of the smoothing, the image
returns to the non-inverted state.
In addition to these there are two variables of process
control: FIRST and SECOND, enabling / disabling the first and
second algorithms. And of course the coordinates of the window,
which smooths out: X1, Y1, X2, Y2.
Attached you will find:
- SGLAV1.Z, SGLAV2.Z two source
File algorithms with comments
- SGLAV2 code file ready to
Uses
- SGL.EXMP example of using
- $ BASHK1 picture - an example.
All right, maybe enough for today. In
further rooms, we might (!) continued the theme of perversion
of the pictures.
____________________________________
