ach integrates
igh-resolution
(Pan) images
d MS and Pan
strate that the
ly higher than
| parking lots,
automatically
better extract
rmation from
nation from
chniques, this
ies of image
extraction into
id. edge-aided
en developed.
or urban road
other object
yad extraction
c images. The
w approach is
multi-spectral
ire extraction,
omatic image
1
the proposed
assification of
ckBird images
image. An
to the pan-
nage. And an
nage to obtain
jn, the binary
o segment the
mage. Then, a
ring algorithm
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B3. Istanbul 2004
MS Image (Infrared, Red, Green, Blue) Pan Image
PANSHARP Fusion
Y
Pan-sharpened Image (Infrared, Green, Blue)
Unsupervised
Classification
/
Edge Detection
Classified Road Image A
Binary Edge Image B
|
Edge-aided Segmentation
Classified Road Image Cut by Edges
Shape-based Segmentation & Segment filtering
Y
Enhanced Classified Road Image <
Y
Final classified Road Image
Edge-aided Segmentation
Figure 1. General process of the proposed image fusion, classification and future extraction integrated approach
The whole edge-aided classification process can be iterated to
deal with complex road classification results. The individual
processes of the proposed approach are described in the
following sections.
2.] Pan-Sharpening
Pan-sharpening is a technique that produces a high-resolution
MS image by combining a low-resolution MS image with a
high-resolution Pan image. The pan-sharpening technique used
in this study is the PANSHARP module (Zhang, 2002) of PCI
Geomatica. The algorithm achieves a maximal spatial detail
increasing and a minimal color distortion (Zhang 2002). After
the fusion, a pan-sharpened QuickBird image is obtained with a
0.61m resolution and 4 MS bands (NIR/R/G/B). Figure 2 shows
a sub-scene of the original QuickBird Pan, MS and the pan-
sharpened images.
M E
i
