veying and
g AIRSAR
is methods.
aging radar
er insights
logy of the
form those
sets. This
en surface
n units. An
na and its
e similar to
| a manner
light and
ttering and
on of high-
veying and
ation.
idar system
it also for
er insights
g AIRSAR
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B2. Istanbul 2t
In the experiment, No.38th research institute. designed and
constructed a high resolution X-band SAR system.
In the following, an outline of the system design is given: the
SAR system is configured as one antenna radar with a ground
resolution up to 0.5m X 0.5m. Figure | shows the AIRSAR
flight segment.
Due to its compact design, the SAR system can be installed on
rather small aircrafts. During the projects mentioned, the system
was installed in Y 12.
3. TEXTURE ANANLYSIS
3.1 High Resolution AIRSAR Images
High resolution sensor technique has made great process since
1990's. High resolution image can show the object information
such as structure, texture and detail clearly. Texture feature is
the direct embodiment of the object structure and space
arrangement in the image. Recently, resolution of remote
sensing image data has been higher, and the tendency has been
seen not only for visible sensor images but also for Synthetic
Aperture Radar (SAR) images. Textural analysis has also been
carried out for SAR. However, it is scarcely discussed that
textural features of high resolution AIRSAR images such as
river, road and residential area are extracted by texture analysis.
In this study, the influences carried by the shadows of
residential area, trees and mountain areas that have been
considered in high resolution AIRSAR. However, unusual
surface patterns are formed clearly due to the fine shadow
included the analysis areas using higher resolution images.
3.2 Texture Analysis
Image texture in general is considered the change and repeat of
image grey in space, or local pattern (texture cell) in image and
its arrange rules. Texture is the important information in remote
sensing and significant base of interpretation by manual work
and computer. In extracting remote sensing image thematic
information, it improves the correction and precise through
adding texture information to original image spectral
information. Image texture, defined as a function of the spatial
variation in pixel intensities (grey values), is useful in a variety
of applications and has been a subject of intense study by many
researchers. One immediate application of image texture is the
recognition of image regions using texture properties.
Applied texture method is to carry out texture analysis. Texture
analysis refers to acquire texture character through some image
processing technology, then obtains quantitative or qualitative
description of texture. It includes two aspects: inspecting basic
cells of texture and acquiring the information on basic cells
arrange distribution of texture. Statistics-based | method,
structure-based method and spectrum-based method are put
forward (Jiang et al, 2003). Statistic method refers to carrying
out texture analysis in the condition of unknown the basic cell
of texture, and it mainly describes the basic cell of texture or
random and spatial statistic character in local pattern, such as
GLCM (Grey Level Co-occurrence Matrices), wave transforms,
fractal representation, *visual" properties random field models
and other representation. Structural texture analysis focuses
primarily on identifying periodicity in texture or on identifying
their placement rules.
161
Texture analysis has been extensively used to classify remote:
sensed images. Filtering features and co-occurrence have been
compared in several studies, which concluded that co-
occurrence features give the best performance. Co-occurrence
technique use spatial grey level difference based statistics to
extract texture from remote-sensed images.
Rignot and Kwok (Rignot et al, 1990) have analyzed SAR
images using texture features computed from gray level co-
occurrence matrices. However, they supplement these features
with knowledge about the properties of SAR images. Du (Du,
1990) used texture features derived from Gabor filters to
segment SAR images. He successfully segmented the
SAR images into categories of water, new forming ice,
older ice, and multi-year ice. Lee and Philpot (Lee et al,
1990) also used spectral texture features to segment SAR
images.
3.3 Grey-level Co-occurrence Matrix
Co-occurrence matrix representation:
I. Method of extracting properties of an image by
comparing grey-tone spatial dependencies between
pixels;
Matrices of the frequencies (probabilities) of going
from one gray level to another at a predefined
distance and different orientations is derived;
3. 14 Statistical measures of texture can be extracted
from the matrix into a feature vector, E.g. Inverse
Difference Moment, Energy (Angular Second
Moment), Contrast, Correlation, Entropy;
t2
Grey-level co-occurrence matrix is the two dimensional matrix
dE & gest tou
of joint probabilities p ; 7 between pairs of pixels,
separated by a distance, d, in a given direction, r. It is popular in
texture description and based on the repeated occurrence of
some grey level configuration in the texture; this configuration
varies rapidly with distance in fine textures, slowly in coarse
textures.
Finding texture features from gray-level co-occurrence matrix
for texture classification in this experiment are based on these
criteria (Mihran, 1998):
Energy:
2,2, P0. uU
gu
Entropy:
SN P, j Mog P, GG. 7) (2)
I
Contrast: (typically k = 2,1= 1)
SG
—
wo
—
———Á—À
PE
