tanbul 2004
(8)
(9)
6 (10)
the loaf can
image and
image and
r and lower
? dominant
component.
from the
al way of
image is to
t the hue
rry out a
g the RGB
| extraction
| with its
)
bread
ground, a
is applied
noise, and
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
non-consistency in the binarized image are eliminated by
carrying out a morphological structuring on the image as
following:
Pita2 := | — binarize(Pital , 60)
Figure 3. Binary image of the pita bread
1:=0..4 j=0..4 msi ; :=0
ms; | zilü-4^5-4Y «19,1,
Pita3 := open (Pita2, ms, 1)
Pita4 := close(Pita3, ms, 1)
Figure 4. Morphologically structured image of the pita bread
From the binarized image, the surface area of the pita loaf can
be measured easily as follows:
nr := rows(Pita4)
nc := cols(Pita4)
nr—1 nc-l
Area := à. >, Pita4; ;
i=0 j=0
Area = 10975
The colour and extension of the dominant colour in the loaf was
determined by applying masking process on the hue saturation
image of the loaf using the binarized image for the mask, then
481
binarizing the resulting masked image using upper and lower
thresholds to define the dominant colour. The extension of the
dominant colour is determined by subtraction of the two
threshold images. The area of the dominant colour and its
percentage compared to the loaf area was calculated according
to the following equations:
UT := 240 LI = 10
PitaS := mask(Pital, Pita4)
ThresholdU := binarize(Pita5, UT)
ThresholdL := binarize(Pita5, LT)
Dominant Colour := ThresholdL — ThresholdU
nr—1 ne-l
Dominant Area := > 3 Dominant Colour; ;
[m0 1-0
Dominant Area — 9609
Percentage of dominant colour:
Dominant Area
Percentage of Dominant Colour := 100
Area
Percentage of Dominant Colour = 87.554
. The average gray scale of the pita bread can be calculated as the
summation of gray scales larger than zero and dividing this sun
by the number of pixels having this "larger than zero" gray
scale, the calculation is done according to the following
function:
Av GS(IMG):s [k —— 0
AV «0
for i € 0..rows(IMG) — 1
for j € 0... cols(IMG) — 1
if IMG, ; >0
AV « AV + IMG; ;
kek+1
AV
return -—
k
Av_GS(Pital) = 66.272
The gray scale and extension of localized colours can be
determined and calculated in the same way as in the case of
dominant colour, where upper and lower thresholds for the
localized colour has to be chosen, then the subimage of the
localized colour is determined by subtracting the lower
threshold subimage from the upper threshold subimage. The
area of the localized colour is calculated in the same way used
in the case of the dominant colour except that the subimage of