BEGINNING.
The theme of this and more of the following articles -
scaling images. We consider the basic methods of scaling
image. All this will be
considered by the example of zooming in twice.
With the advent of the market file
Image IBM standards, and
otsifrovok images before
I faced the problem of reducing these pictures to screen size.
It is most urgent, for the owners of 48 KB
computers that you want
look at not reaching them
pictures, as well as artists and
people who want a good draw
their program.
On a floppy disk you'll be offered an example of a program
to reduce image files with any extension of twice.
Wesy.asm program designed to give a universal base for your
own further research this area.
BASIC METHODS
Scaling.
1) Delete / add extra
pixels;
2) Averaging and metal threshold
method:
a) uniform;
b) The weighted;
3) approximation of semitones;
Removing unnecessary pixels.
The method is based on the fact
that a certain number of pixels in the original image
zoom out (the compression of the image) is simply not included
in the resulting image. Or, if there magnification, what
then the original screen pixels are duplicated.
This algorithm is more efficient to reduce image
to 1,5 times, but sometimes halving has a good
result.
Reduction algorithm in two times is simple: Take a starting
point (Eg (0,0)) and copy it
on the screen at (1,1), go to
the next point (2,0), copy on
(1,0), etc. to fill the screen.
The program of this method is implemented on a label BEIN2.
Disadvantages of the method: when you remove half of the
information is lost and simply ignored.
Averaging.
So we have a compressible
Image: 512 * 384. We want
reduce it by half and get a display file 256 * 192.
We see that each
system on display pixel corresponds to 4 points of the original
image (2 * 2). And are based on this the following methods.
Averaging can be of two types.
With a uniform averaging we
count the number of pixels with
Color INK. The amount will not be
greater than 4. This amount is called
intensity of display pixel I. If the sum is greater than 2, then
a display pixel has a color
INK, otherwise PAPER. Thus
determine the color at each point
screen.
Location Showing
pixel on the screen
A B I = int ((A + B + C + D) / 2.5)
C D
Here A, B, C, D-color pixel at the corresponding point of a
compressible image. (INK-1, PAPER-0);
You can get better results if we use
method of weighted averaging.
This method takes into account the greater number of points
(not 4), assigning each point of your weight. (cm
2). Then the resulting intensity will be:
I_R = 1 * A +2 * B +1 * C +2 * D +4 * E +2 * F +
+1 * G +2 * H +1 * K, where
A, B, C, D, E, F, G, H, K-pixel color in
corresponding point on the compressed
image. (INK-1, PAPER-0);
E - central pixel in the analysis.
LOCATION WEIGHT
pixel pixel
A B C 1 2 1
D E F 2 4 2
G H K 1 2 1
Figure 2. Weighted averaging.
Calculated the resulting
intensity we compare with
number 8. Draw color INK point on the screen, if I_R more than
8. Then go to the next coordinate (increasing the current
coordinates on the 2 on one of the axes
and if you want the second). That is
we make the central point
compressed image through one.
Sub-weighted averaging is located on a label
BEGIN.
What makes for "Spectrum" threshold methods from methods
averaging? Virtually nothing.
We simply compare the resulting
intensity of the display pixel is not a fixed number (2
or 8), a with a certain threshold
value of the intensity I, entered from the keyboard.
TO BE CONTINUED
; Works CLS (without RESET POINT)
, Ie CLS needed before starting
; Weighting method
START JR BEGIN
NOP
; Removal Methods
START2 JP BEIN2
DEFS 27
; SUB file download
; CD
LOADER JP LOAD
; Image intensity
INTENSIVE
DEFB 5
; To what coord. ON X PRINT
; In pixels Compress files.
RAZM_X DEFW 376
; To what coord. PRINT ON Y
RAZM_Y DEFW 270
, The number of bytes into a string
; IMAGES
RAZ_B DEFW 82
; Start address Compress files
NACALO DEFW 30018
, At which point Compress files
; BEGINS WORK ALGORITHM
NAC_X DEFW 0001
NAC_Y DEFW 0001
; WHERE loaded in Compressible Files
LOA_BE DEFW 30000
; END describe the data.
; BALANCING ALGORITHM
BEGIN
LD DE, (NAC_X)
; X-0 .. 255
CIKL_I LD BC, (NAC_Y)
; Y-0 .. 176
; Zero the H.
, Formed in the resultant
, The intensity of the analyzed
; Pixels compressed image
CIKL_J LD H, 0
; Save the coordinates:
; X (DE) and Y (BC)
PUSH DE
PUSH BC
; Next, we analyze the point of contraction; ible images in the
following ; Order: have weights:
; A B C 1 2 1
; D E F 2 4 2
; G H K 1 2 1
;
; *** Analysis of point A:
, Shifted by A
DEC DE
DEC BC
; Analyze the value of a point on
; According to the coordinates X (DE) and Y (BC)
; The output of 0 or 1.
CALL POINT
; Add the value of the pixel to
, The resulting efficiency.
ADD A, H
LD H, A
; *** Analysis of point B
; Shifts to B
INC DE
CALL POINT
; Of gaining intensity, T at this point, the image the mind;
knife at 2. See matrix.
SLA A
ADD A, H
LD H, A
; *** Analysis of point C
INC DE
CALL POINT
ADD A, H
LD H, A
; *** Analysis of F
INC BC
CALL POINT
SLA A
ADD A, H
LD H, A
; *** Analysis of point E
DEC DE
CALL POINT
SLA A
SLA A
ADD A, H
LD H, A
; *** Analysis of point D
DEC DE
CALL POINT
SLA A
ADD A, H
LD H, A
; *** Analysis of G-Spot
INC BC
CALL POINT
ADD A, H
LD H, A
; *** Analysis of the H-point
INC DE
CALL POINT
SLA A
ADD A, H
LD H, A
; *** Analysis of the K point
INC DE
CALL POINT
ADD A, H
LD H, A
; H = A + C + G + K +2 * (D + B + F + H) +4 * E
, Determines the coordinates of O
, A point on the screen
DEC DE
DEC BC
SRA D
RR E
SRA B
RR C
; Compare the intensity of the problem, under consideration,
the user account, and Noah.
LD A, (INTENSIVE)
CP H
JR NC, NEXT
; If INTENSIVE-H <0, then draw
; Point on the screen. If not, then
; Circumvent routine painting.
CALL SETING
; To recall the old X and Y
NEXT POP BC
POP DE
; Transition into the next Y and verified; ka at the end of
vertical construction, the vertical line.
INC BC
INC BC
LD HL, (RAZM_Y)
SBC HL, BC
JP NC, CIKL_J
; If pressed any Clavey; step, the output from the subroutine
BIT 5, (IY +1)
RET NZ
; Transition into the next X and verified; ka at the end of the
construction of the screen.
INC DE
INC DE
LD HL, (RAZM_X)
SBC HL, DE
JP NC, CIKL_I
RET
; *** END OF THE WEIGHT OF THE METHOD
; *** Deletion algorithm
BEIN2
; Download in DE and BC are the initial
, The values of X and Y.
LD DE, (NAC_X)
; X-0 .. 255
C2KL_I LD BC, (NAC_Y)
; Y-0 .. 176
C2KL_J
; Memorization coordinates X, Y.
PUSH DE
PUSH BC
; Processing point coordinates.
CALL POINT
And displays schitanye point on
; Screen.
SRA D
RR E
SRA B
RR C
AND A
JR Z, NEXT2
CALL SETING
; Remembering X and Y.
NEXT2 POP BC
POP DE
INC BC
INC BC
; Go to the following coordinates
LD HL, (RAZM_Y)
SBC HL, BC
JP NC, C2KL_J
INC DE
INC DE
LD HL, (RAZM_X)
SBC HL, DE
JP NC, C2KL_I
RET
; *** End of removal.
; *** Installing point
; To screen coordinates:
; BC-Y (upper left)
; DE-X
SETING PUSH BC
PUSH DE
; Address the pixel count
, And the number of bits.
CALL RASCHET
; Converts byte mask
LD B, # 0FF
XOR B
OR (HL)
LD (HL), A
POP DE
POP BC
RET
; *** End of installation procedures
; *** Point.
; *** The procedure of analysis point in
; Compresses the images.
POINT PUSH BC
PUSH DE
PUSH HL
; Get the address in the file is compressed; ICDO image and
select ; The desired bit.
CALL RASHET
LD B, # 0FF
XOR B
LD B, A
LD A, (HL)
AND B
, Check to see if the desired bit 0
; If not, then send to A = 1,
, Otherwise it remains 0
JR Z, NEXT_POINT
LD A, 1
NEXT_POINT *
; Remembering all the registers.
POP HL
POP DE
POP BC
RET
; *** End of analysis procedures.
; *** The procedure for calculating the address
; Familiarity in a compressible file.
RASHET
; Memorize X
PUSH DE
; In HL address of the first Bait
; Image.
; Believe HL = HL + DE / 8
LD HL, (NACALO)
SRA D
RR E
SRA D
RR E
SRA D
RR E
ADD HL, DE
; Read the size of a single row
; Image.
LD DE, (RAZ_B)
; HL = HL + DE * BC
SRAD SRA D
RR E
JR NC, NOL
ADD HL, BC
NOL SLA C
RL B
LD A, D
OR E
JR NZ, SRAD
; In HL-ADDRESS. It remains to find bits.
END1 POP DE
LD A, E
AND # 7
INC A
LD B, A
LD A, # FE
LOOPA RRCA
DJNZ LOOPA
, NOW IN A will be 0 in the right
; Discharge.
RET
; *** End of calculating an address for
; POINT
; *** Calculating addresses for SETTING
; OUTPUT IN (HL)-byte at
; In B 11101111 IN THE APPROPRIATE
; Discharge. For more details see
, Applied Graphics, Inforkom
RASCHET
LD A, E
LD D, A
LD A, C
LD E, A
AND A
RRA
SCF
RRA
AND A
RRA
XOR E
AND # 0F8
XOR E
LD H, A
LD A, D
RLCA
RLCA
RLCA
XOR E
AND # C7
XOR E
RLCA
RLCA
LD L, A
LD A, D
AND 7
LD B, A
INC B
LD A, # FE
LOOP RRCA
DJNZ LOOP
RET
; *** End of sub SETTING
; *** Sub-boot.
LOAD
; Adjusts to find a file on
; Name (8 characters).
LD A, 8
LD (23814), A
; We are looking for a file named stored on
, At 23773.
LD BC, 10
CALL # 3D13
; Check if the file in the kata; log.
LD A, C
BIT 7, A
RET NZ
; Read the file name.
LD C, 8
CALL # 3D13
; Read the file found
LD HL, (LOA_BEG)
LD A, (23786)
LD DE, (23787)
LD B, A
LD C, 5
CALL # 3D13
RET
; *** End of sub downloads.
BASIC PROGRAM.
On the disc are two demo BASIC program:
WesyScr - for a reduction in
double screen
files.
Scma.02 - for file format
IMAGE.
Code blocks these programs
slightly different.
To be halved
Zoom in Scma.02 you need
to know its format. To do this,
look in the "IBM SCREENER" this screen, set it up the keys 9
and 0, and press 1. Next will give you a hint. Remember her
from the number of bytes in line. Next, read the file
Scma.02, a picture and a question
"LINE SIZE?" enter that number.
If you press "ENTER" will
taken the previous number (in the
early 82).
Next you are asked about the form of the algorithm: 0 -
gravimetric method
1 - removal method
Enter the intensity and the initial pixel processing
(0th or first X and Y).
If you want to cut some
another piece of the image,
ispravte starting address of 100
line.
After the construction of (or in the process of building for
the gravimetric method), pressing the following keys will
result in any effect:
S - written on the current screen
disc
E - go to BASIC.
L - switch to input the following items
file.
I - to introduce a threshold intensity
The intensity and begin to construct
enie again.
Any other key will result in
change all the parameters of the program.
