th

The 14 World Conference on Earthquake Engineering

October 12-17, 2008, Beijing, China

Study on Digital Earthquake-before Damage Evaluation

1

Min Zhu Zhongke Feng Huijie Tong Yongxia Liu Yankai Yang
1

Associate Researcher, Institute of Engineering Mechanics, Harbin. China

2
Professor, Beijing Forestry University ,Beijing, China
3
Associate Professor, Beijing Forestry University ,Beijing, China
4
Student Researcher, Institute of Engineering Mechanics, Harbin. China
Email: [email protected], [email protected]

ABSTRACT :
Buildings digital auxiliary classification, summation of different kinds of buildings area, for earthquake-before
damage evaluation investigation, by integrated application of RS (Remote Sensing) and DPS (Digital
Photogrammetry System) and otherwise digital method are studied. Two kinds of buildings digital auxiliary
classification methods for earthquake-before damage evaluation have been used. If there are remote sensing
image stereo pairs in the evaluated area, DPS can be used. A number of DSM (digital surface models) of
three-dimensional structures and DEM can be obtained. There is another earthquake-before damage evaluation
survey method that buildings height are calculated automatically by buildings shadows in remote sensing image.
Buildings can be aided classified by height. According to formulas of aerial survey, buildings height can be
calculated by the shadows. The buildings shadows can be picked up by a series of RS image transformations.
The area summation of different kind of buildings can be obtained by GIS. Digital earthquake-before damage
evaluation, which combines digital technology with experience and statistical evaluation, is faster than
traditional methods.
KEYWORDS: Digital classification, Earthquake-before damage evaluation, DPS, RS

1. INTRODUCTION
Digital technology has been applied to prevention and reduction disaster in recent years. RS and DPS, two
kinds of digital technology have been applied to earthquake before damage evaluation and environment change
measurement.
Narpat Singh Rathore utilized RS, to study topography change earthquake bringing. Nioki Ogawa and
Hirotada Hasegawa estimated respectively earthquake damage of wood buildings by Kobe earthquake aerial
remote sensing images. Some micro vertical diastrophism of active fault was monitored and surveyed utilizing
radar images. Earthquake tectonic maps of remote sensing images interpretation in china have been compiled
by seismologists and rocksies and earthquake engineering scientists in the sevens of the twentieth century.
Aerial remote sensing images were interpreted artifically by Decheng Zhang for earthquake damage evaluation
of buildings in 1993. The aerial remote sensing images were utilized by Yang Zhe to investigation of
Lanchang-Gengma earthquake disaster. Earthquake damage evaluation of disaster area was studied by Zhang
Jingfa contrasting radar remote sensing earthquake stricken before with after.
Change of topographic and wild sight was studied by Gong Peng with DPS. Performance of concrete
girder bearing shearing force was studied by Qu Zhe with DPS.
Automatic extraction of shadows from high resolution spatial satellite images were studied by Xu
Miaozhong. The distribution information of different heights building in a city was extracted from the shadows
in an Ikonos image by Xie Junfei.

th

The 14 World Conference on Earthquake Engineering

October 12-17, 2008, Beijing, China

Earthquake before damage evaluation in china began in the sevens last century after Tangshan earthquake.
The earthquake vulnerability analysis is the primary means for earthquake before damage evaluation. The
structure will be destroyed if outside force put is over a threshold, this can be evaluated and calculated by a
number of parameters on type of buildings and building materials and so on. First of all works for earthquake
before damage evaluation, the buildings are classified according to the principles of vulnerability classification.
The principles of vulnerability categories are according to earthquake damage with similar buildings in the past
earthquakes. The field survey of the evaluation area was the basis for earthquake vulnerability calculation.
Sample buildings were surveyed detailedly for statistical calculation of vulnerability. The more sample
buildings the better the result of statistical calculation is.
In traditional field survey of earthquake before damage evaluation, a number of representative buildings
chosen from each kind of general-purpose buildings were surveyed one by one. A series of information for
earthquake before damage evaluation can be obtained by calculating with results surveyed. A lot of field survey
information can be obtained by digital technology. People on the ground of limited sight, are lack of overall
macro distribution of different type buildings. The buildings are overlooked in the air utilizing RS and DPS.
There are characteristics of overlay area extensive and evaluation fast. RS and DPS digital technology were
applied to aid classification on field survey for earthquake before damage evaluation.

2. EARTHQUAKE BEFORE DAMAGE EVALUATION WITH DPS

Pairs of aerial anaglyphs in Beijing were utilized for digital test. DPS software was utilized for relative
orientation and absolute orientation and matching editing, DSM and DEM were constructed. The height of
buildings was obtained automatically. Buildings were classified by height, assisting survey for earthquake
before damage evaluation.
After setting parameters and survey area being set up, the model was set up and orientated. Nuclear-line
was resampled and Nuclear-line image was obtained. A part of three-dimensional model of buildings is shown
in figure 1.Buildings were measured by digital measure of DPS. Finally, the digital map with three-dimensional
coordinates of buildings was obtained in the end and is shown in figure 2.

Figure 1 three-dimensional model of buildings

Figure2 results of the digital mapping

More than 600 buildings were digital measured in the digital test. The information of DSM and DEM and
the types of building roof were obtained. The height of a building can be obtained by formula 2.1.
h = DSM-DEM
(2.1)
In formula 2.1, h is a building height, DSM is a value of digital surface model and DEM is a value of

th

The 14 World Conference on Earthquake Engineering

October 12-17, 2008, Beijing, China

digital elevation model.

According to the height of the building and each floor, the number of floors was calculated. A total building
area was calculated by the plane area and the number of floors. The digital test results are shown in table 2.1.
Table 2.1 Results of digital faster buildings classification in test area
Cottage
Multi-storey building
High building
2
2
Number Total area(m ) Number Total area (m ) Number Total area (m 2 )
203
33325.994
354
947950.175
92
1465983.383

3.EARTHQUAKE-BEFORE DAMAGE EVALUATION WITH RS AND SHADOW

The information on drop shadow of buildings can be extracted automatically by RS. The building height
was calculated with drop shadow. Assisted classification for earthquake before damage evaluation with RS and
building shadow was studied.
Processed with geometric correction and enhancement and so on, the remote sensing image was classified,
unsupervised and supervised. Part of unsupervised classification results are shown in figure 3.

Figure4 Partial supervised classification results

Figure 3 Partial unsupervised classification results

Blue represented drop shadow of buildings in figure 3. Other kinds of color represented different kind
of buildings roof , a part of buildings , road, trees and so on. There were some confusion categories
between shadow and water, buildings and roads. It can be improved by aided information.

After unsupervised classifying, supervised classification is utilized. Trees, buildings with five color roof,
roads and drop shadow are included in supervised classification template. The possibility matrix was utilized to
estimate classification template and it is shown in table 3.1.
Table 3.1
Result Temple

Possibility matrix to estimate classification template

Tree

East-west

Grey roof

Brown roof

Light-grey roof

White roof

Road

Shadow

Tree

11276

12

10

162

East-west

30

964

509

48

Objects

th

The 14 World Conference on Earthquake Engineering

October 12-17, 2008, Beijing, China

Grey roof

4628

61

Brown roof

528

Light-grey roof

12

17

991

White roof

30

740

36

Road

465

51

550

Shadow

445

4685

Total

11770

994

5652

528

1052

852

589

4847

The classification results are more satisfactory. There are some confused category to be improved between
roads and buildings. The results need to be improved through referring to special topic and field survey
information. Partial supervised category results are shown in figure 4.
In figure 4, green represented trees. Blue, grey, yellow, purple and white represented different color
roof of buildings. Black represented drop shadow of buildings.

According to aerial survey formulae, buildings height is calculated with buildings shadow. The length of
buildings shadow extracted automatically by RS, as shown in figure 3, was calculated automatically by GIS.
Shadow of the buildings were measured, the partial results are shown in table 3.2.
Table 3.2 Buildings shadow length calculated automatically
Number Shadow length(m) Number Shadow length (m) Number Shadow length (m)
10
20.8173
19
13.0384
28
17.366635
11
23.021729
20
8.485281
29
28.319605
12
23.64148
21
3.622154
30
19.284
13
18.508376
22
17.075128
31
8.551023
14
9.338094
23
34.567615
32
8.318654
15
8.551023
24
4.701644
33
22.951253
16
15.770859
25
4.837355
34
16.643317
17
6.003
26
7.002857
35
5.60357
18
13.0384
27
9.535198
36
22.598521
According to shadow as above, building height and floor number are calculated. The results are shown in
table 3.3.
Table 3.3 Buildings height and floor number
Number Height(m) Floor number
Opinion Number Height(m) Floor number
10
22.212
7
right
24
5.0166
1
11
24.5641
8
right
25
5.1614
1
12
25.2254
8
right
26
7.472
2
13
19.7484
6
right
27
10.174
3
14
9.9637
3
right
28
18.5302
6
15
9.1239
3
right
29
30.217
10
16
16.8275
5
right
30
20.576
6
17
6.4052
2
right
31
9.1239
3
18
13.9119
4
right
32
8.876
2

Opinion
right
right
right
right
right
right
right
right
right

th

The 14 World Conference on Earthquake Engineering

October 12-17, 2008, Beijing, China

19
20
21
22
23

13.9119
9.0538
3.8648
18.2191
36.8836

4
right
3
right
1 low value
6
right
12
right

33
34
35
36

24.4889
17.7584
5.979
24.1126

8
right
5
right
2 more low
8
right

Comparison results between digital and actual classification, the accurate rate is 80%.The accurate rate of
classification for low-rise and smaller buildings is not ideal because the buildings distribution was denser and
there were a lot of high-rise buildings in the test area.

4. CONCLUSION
The digital technology can play an important part in earthquake damage evaluation when it is more dificult
to field survey. The RS and DPS can be utilized to quantitative survey of wide scope in a short period of time by
interpreting high image. The results can covered a broad area and be more precision. To improve the efficiency
of earthquake before damage evaluation using digital technology is a very significant. It is great importance to
development of prevention and reduction earthquake disaster.
REFERENCES
Nioki Ogawa,Furmio Yamazaki.(2000),Photo-interpretation of building damage due to earthquakes using aerial
photographs,Proc of 12 WCEE,Vol 15:1906,New Zealand Society for Earthquake Engineering.
Zhang Decheng.(1993), Preliminary study on visual interpretation marks of building damages caused by
earthquakes on aerophotograph. Earthquake 1, 26-30.
Zhang Jingfa, Xie Lili and Tao Xiaxin (2001).Pattern analysis of remote sensing imagery for some typical
earthquake damages. Journal of Natural Disasters 10:2, 89-95.
Gong Peng, Mei Xieliang and Gregory S. Biging (1999). Monitoring oak woodland change using digital
photogrammetry. Journal of remote sensing 3:4, 285-288.
Gong Peng (2000).Digital surface model and topographic change monitoring. Quaternary sciences 20:3,
247-251.