tion in multi-
re intensive
3)
1e of images
similarity of
better fusion.
(4)
mages before
e mean value
CM, they are
ween them is
erent moving
iral measures
| measures of
ame linear or
| — (g)). The
dow size of
Homogeneity
n negative to
ist with other
ncy, but the
ogeneity and
vindow sizes,
respectively.
ved the same
'e correlation
| others all
wed that the
; between the
ndancy.
of an image,
om Fig.2 (h)
creasing with
at the size of
n to calculate
as applied to
o replace the
ectral image.
by GLCM, and correlations between them, which has been analyzed
above. The textural measure of Entropy based fusion has more spatial
information, and less spectral distortion. Homogeneity based fusion
has the least information, and Second moment based fusion has the
largest Spectral distortions. So when we fuse textural measures with
images, it is important to choose the textural measures.
Then the evaluation index, Shannon entropy, sharpness, spectral
distortion and correlation coefficient were applied to evaluate the
fusion results. From Table 1, we can see that the information
abundance of fusion with single textural measure are all increased.
That means it is reasonable to fuse texture with images. But the
results are close to each other. That is because they are all obtained
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B7, 2012
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia
"exture measures Sharpness Shannon Spectral correlation
entropy distortions coefficient
Mean 7.3447 1.1949 25.0928 0.7536
Variance 6.2519 1.1270 24.5272 0.7387
Homogeneity 6.9448 0.9866 27.9070 0.7809
Contrast 6.2741 1.1281 24.4223 0.7399
Dissimilarity 7.5511 1.1983 24.6145 0.7624
Entropy 8.2487 1.2574 24.9710 0.7811
Second moment 8.2528 1.0632 33.3779 0.7151
Original QuickBird image 10.6940 0.9332 --- -—
Table 1. Evaluation of PCA fusion based on texture analysis
5. CONCLUSION
The results of the procedure proposed in this letter show that it is
possible to extract some kind of texture information high satellite
SAR images and fuse them with some methods. In this study, texture
has been extracted by GLCM after analysis, and fusion performed
simply by PCA at pixel level. The results are not so satisfactory. That
[^ qvYTICYYCETEY * + *
+
: d
set ilzaae doo De
ah sai ac RM vod
; coves tittle + *
Gs as
té €
2 sat Ss
E E
¥ ia s
egt S
3 s Currelatiu of Mean = *
3 on d ; + i 3. i =
M 28 FF à m Bk Fe: yp Bisintant
= te 2 s y > }
X. E "i A 3 Correlation af Variance E
w GEE s xa* Window Site > 3 Seonnd Moment
x* 5 2n + + 3 + + hy i
$a x a s 1 BR AR 7
E * sent
» pe Sau at Window Size
„1 28 war: £2 it.
QbRatjésqa oo 3J*ex43443 4 + >
QU (b)
ir 380000008 * » Wr .... . eod
asas „se ........
*
nll ESL = 3 3
£ : > EFFRITQUTEEEE 5
a N rd
* Sek
€ 3
325k 3 24
: © 8
É €
à E % 3 atn
$ 24 $39 Correlatiun of Déssimilarity
=
5 i. ig Pa
= 3 + 4 t
Bu : Wo wq c ES 284
ET Correlation of Canteast 3 CU
= orcelatiog of Canteas 348 Vii si
© * NN iut © Jado She
50 i i 1 x i reed En à 4 *
= “ » Li A [s Te $i *
i E "mI
E SI Window Size N
; PH 4 + * AB
(d) (e)
ssh *
E. t à
Eo +
3 82) — Correlation of Second Moawat T
S
3 Windo $ =
Snap 8$ w Sire
8 | eee queas
“sh 2 id
^
$8 Psegponeut 9 à
re
â a
Qmm
42
Ntsestard Des tation
x
Tsnoussans
Cerretation Coefficirnt
Correlation Coelficient
t6 x
is because GLCM is based on gray value statistics, and SAR image
after filtering is inevitably blurred. Some texture information would
be removed. In the following work, other methods (like Wavelet
based) to extract texture should be proposed to improve to obtain
accuracy texture, and other fusion methods can also be used.
x
*
“er,
i Bow
*
*
T 14
BEE
of!
ES
E Medi s $33
x 3 p
* imos Sire
(h)
* + + s:
RE EN
£8
esi
BZ
té A
22 Ub Muss
> 9e Vaname
U4 i Contes
*oc Dsssiilsnity
iR. Varo
+ Correlation of Homogeneity M
2 57 Sezotil Meunzcat
48
(c)
té
a: A P
ax b EL inii
E
x
she" : +
a
Qn
34 Fl
GC
a:ba*
ga + eed. t + +
" wa x au ie &
~ ET x, = *
83H A Window Size *
Sa : i »
4#* S 3
» &
3 at s
* +
A8
25 = Corvefation of Émropy
* *
tk
