data collection and inventory compilation methods for seismic risk assessment … · 2014. 9....

Systemic Seismic Vulnerability and Risk Analysis

for Buildings, Lifeline Networks and

Infrastructures Safety Gain

Data collection and inventory compilation

methods for seismic risk assessment:

Application of remote sensing techniques

for building inventory update

Ufuk Hancilar1, Patrizia Tenerelli2, Fabio Taucer3, Daniele Ehrlich3,

Sotiris Argyroudis4, Kyriazis Pitilakis4

1 Dept. of Earthquake Engineering, Boğaziçi University, Istanbul

2 IRSTEA Centre de Grenoble, UR Ecosystèmes Montagnards, France 3 Joint Research Centre of the European Commission (EC- JRC), Ispra 4 Lab. of Soil Mec., Foundations & Geotech. Earthquake Eng., Aristotle

University, Thessaloniki

2

Structures, utilities, systems and, population and socio-economic activities constitute the “Elements at Risk” in urban areas.

The physical elements are the built environment such as buildings, lifeline networks, transportation infrastructures, etc., while the social elements are represented by the demographic and socio-economic data.

It is an essential step in urban earthquake risk assessment to compile inventory databases of elements at risk and to make a classification on the basis of pre-defined typology definitions.

Typology definitions and the classification system should reflect the vulnerability characteristics of the exposed inventory in order to ensure a uniform interpretation of data and risk analyses results.

Introduction

Introduction

SYNER-G considers four main categories of systems:

1. Buildings: Reinforced Concrete and Masonry

2. Utility Networks: Water, Waste Water, Gas, Oil, and Electricity

3. Transportation Infrastructures: Roadways, Railways, Bridges and Harbour systems

4. Critical Facilities: Health-Care and Fire-Fighting Facilities

Introduction

Data sources and collection methods for the compilation of inventories for the purpose of seismic risk assessment can be categorised into four groups:

1. Census and owner/operator data

2. Ground surveys

3. Remote sensing techniques

4. Crowd sourcing

Remote sensing data types and detectable physical parameters Typical European elements

at risk

Visible from

remote sensing

Automatic and semi-

automatic detection

Physical parameter that

can be identified

Suggested data

types

BUILDINGS Yes Possible Building location, planar view,

building density, height, volume, roof type, age

Optical VHR, HR, MR; Stereo HR; Hyper-spectral; Oblique

Aerial; LIDAR; SAR U

TIL

ITY

NE

TW

OR

KS

Electric power

system Partially Not possible

Location and geometric parameters of elements above the earth surface (i.e. power

stations)

Optical VHR; Stereo HR; Oblique Aerial;

LIDAR

Gas and oil

network Partially Not possible

Location and geometric parameters of elements above the earth surface (i.e. pipelines,

tanks)


LIDAR

Water and

waste-water

system

Partially Not possible

Location and geometric parameters of elements above the earth surface (i.e. pipelines,

dams)


LIDAR

TR

AN

SP

OR

TA

TIO

N

INF

RA

ST

RU

CT

UR

ES

Roadway

bridges Partially Possible Bridge location, width


LIDAR

Roadway

system Yes Possible Road main axe and road width

Optical VHR; SAR; Hyperspectral

Railway system Yes Not possible Railway main axis and road width Optical VHR; SAR;

Hyperspectral

Harbour

elements Yes Not possible

Location and geometric parameters of harbour buildings

and cranes


LIDAR

CR

ITIC

AL

FA

CIL

ITIE

S

Health-care

facilities

Partially: secondary

information are necessary to identify the building use

Not possible Location and geometric parameters of facility building


LIDAR

Fire-fighting

system Yes Not possible Location and geometric

parameters of facility building


LIDAR

Existing Building Inventory of Thessaloniki

SYNER-G Thessaloniki, 14-15/6/2012

The existing inventory contains 5,047 buildings out of 19,000 buildings in the municipality. This inventory is based on a combination of the 1991 census data, from the Statistics Agency of Greece, with data collected in a previous project (Penelis et al. 1988) through an in situ survey for 5,047 buildings in 470 blocks following the 1978 earthquake.

Remote Sensing Imagery for the City


A GeoEye-1 image (VHR multispectral image) which covers about 16.5 km2 was available for the analysis. The image was collected in March 2010 with a resolution of 0.5 meter for the panchromatic band (black & white) and 2 meter for the multispectral bands (blue, green, red, near IR).

The panchromatic band was used to enhance the spatial resolution of

the multi-spectral bands using the Gran Smith pan-sharpening technique.

Optical remote sensing application: The case

study of Thessaloniki


Building count • Objective: update the existing footprint location map and

estimate the number of buildings within each block in study area

• Methodology: photo-interpretation of the pan-sharpened GeoEye image for the blocks where:

• information is missing • footprint geometry needs

to be updated • there is mismatch between

mapped building and the building position on the satellite image


Criteria to define a building unit: • homogeneous roof colour • visible distance from the next building • visible difference in the building height

Limitation: • The mapping of single buildings may differ when the

inventory is performed from the ground!


The building count can be used to stratify the building typologies when the typology information is available as

statistical aggregation at the level of building blocks

Building count map




Building area • Objective: automatic extraction of the built-up area • Metodology: texture based algorithm

Map of

Built-up

index

• Validation:

• The final built-up index was compared with the reference building footprint map (1692 blocks)

• The area was compared at 1 m resolution

Overall Accuracy (%) 70.30

Kappa Coefficient 0.34

Confidence level (%) 95

Class

Accuracy (%)

Producer User

Non built-up 76.22 78.95

Built-up 58.11 54.25

Total 67.16 66.60




Building height

• Objective: automatic building height extraction

• Methodology :

1. Building height extraction based on: • length of the casted shadow • satellite viewing angle • sun elevation at the acquisition time

2. Extraction of the height index values on the given building centroids (at 50m resolution)

3. Classification of building height values into storey nr. 4. Aggregation at the building block level

• Final map of number of storeys

• Validation:

• The estimated values were compared with the reference data

• The average storey number was validated for 1692 blocks with: area > 2500 m2 • The low correlation is

mainly due to the sensor parameter limitation

• The best results were found for isolated buildings, where casted shadows are fully visible

Parameter value

Observations (blocks) 1692 Multiple R 0.60

R Square 0.37

RMSE 1.67 F 973

p-value 4.4719E-169

• Objective:

perform spatial metrics and indicators for some built-up layers derived from remote sensing

• Methodology:

• The SHAPE index was applied as a measure of the building shape complexity

• The NEAR index was applied for proximities analysis of building and building to roads

(McGarigal and Marks, 1995)

Combining Remote sensing and GIS

SHAPE index map

• The SHAPE index equals 1 when buildings are compact to the maximum possible extent and increases without limit as the shape becomes more irregular

For SHAPE > 1.5 : complex building footprint shapes

• “T” shape • “L” shape • very elongated

NEAR index map: building to building

• The NEAR index increases as the neighbourhood is increasingly occupied by other buildings and as those become closer and more contiguously distributed

• It can be used to measure the isolation of a building or the distance to the nearest building characterized by a given typology

• The building stock can be estimated from remote sensing data, however structural and functional elements of buildings and civil engineering works cannot be distinguished

• Semi-automatic mapping produces information which may not be enough detailed for operational use at the local level

• The input data types (accuracy, technical parameters), highly affect the final accuracy:

The most suitable remote sensing data to derive building height are very high resolution stereo imagery, or LIDAR data which allow processing 3D surfaces, but have high acquisition costs

Concluding Remarks

19

• Spatial metrics can be extracted and stored as attributes in a GIS vector file and can be used in risk assessment applications

• Land use classifications extracted from remote sensing can be used to stratify information which are available at larger spatial units (building blocks, administrative units)

• Downscaling techniques can be applied to refine the geographical distribution of census dataset, the refined spatial detail can be exploited for exposure and vulnerability analysis

Concluding Remarks

data collection and inventory compilation methods for seismic risk assessment … · 2014. 9....

Documents