! 2004
RESEARCH THE CHANGE OF THE BUILDINGS BEFORE AND AFTER
EARTHQUAKE BY USING AIRBORNE PHOTOGRAPHS
Y. Sidar", B. Bayram®, C. Helvaci®, H. Bayraktar”*
*YTU, Civil Engineering Faculty, 34349 Besiktas Istanbul, Turkey - (bbayram, chelvaci, hbayrak)@yildiz.edu.tr
? Ugur Harita Muh. Burosu, Alemdag Cad. No:210/1-B, Umraniye Istanbul, Turkey- ysidar@yahoo.com
Commission VII, WG VII/5
KEY WORDS: Earthquake, Change, Detection, Photogrammetry, Geometry, Statistics
ABSTRACT:
Recently the possibility of an earthquake has become one of the most important subjects in lots of countries especially in Turkey.
For that reason, the possibility of an earthquake gains ground due to its devastating physical and psychological effects. Because of
these physical losses and its psychological effects, people who experience the earthquake generally become sensitive and provident.
Although disaster management had been established before the earthquake, the studies after the earthquake prove the fact that
decreases in the loss of lives is inadequate. In this study, the aerial photographs have used which have been obtained before and after
the earthquake. An algorithm is developed that determines the damaged buildings as semi-automatic. The principle of this study is
analyzing the buildings’ image histograms and determining their conditions. Gólcük and Adapazan regions before and after the
earthquake have been used for the analysis of the image histograms that belong to the buildings in the aerial photographs. Our
empirical results have shown that the histograms of the undamaged and damaged buildings have different forms. For this reason,
utilizing the histogram analysis, the conditions of the damaged buildings have determined by using a semi-automatic process. The
important pre-condition for this study is the digital maps of the study area to be in existence. Briefly, to form the histograms, the
coverage areas of the buildings that are after the earthquake photographs in vector format are found and the image histograms are
calculated from those areas. The purpose is to determine the damaged buildings quickly during the earthquake and contribute to the
disaster management. It is thought that, through the results that are to be obtained, it will be possible to direct the rescue mission
correctly, estimate the number of loss of lives and take healthy precautions.
1. INTRODUCTION interpretation of aerial images where we want to assess how
much specific algorithms and/or features aid in the completion
The result of development of the computer technology, the of vision tasks. In such vision tasks as target detection, battle
problems of the photogrammetry and remote sensing are carried damage assessment, delineation of buildings and man- made
to vector data. Due to the accessibility of advanced software objects, the segmentation map represents an intermediate level
and hardware systems, it is possible to do computer based intelligence to the human operators (Letournel, 2002).
photogrammetric process. Also, it is feasible to detect the After natural disasters happen, the damages have to be
changes through the development of the satellite technology. determined and an effective disaster management has to be
Monitoring the change is meant to present reliable, quick and conducted in a short time. This was observed to be a very
up-to-date data relating to the analyzed area to the deciders. important problem in the earthquake of 17 August 1999,
(Gunduz, 1996). Systematic and effective manipulation of Kocaeli. (Bayram, 2001).
remotely sensed data is necessary for the utilization of data In this study, aerial photographs which are taken before and
from resource satellites and low-level airborne platforms for after the earthquake are used. The algorithm and application
monitoring and analyzing the dynamic phenomena of the which determine the deformation of the buildings after the
Earth’s surface. Because data capture is sequential over time, earthquake are explained with change detection.
multiple images for any particular area contain potential
information about changes of surface features or conditions. 2. APPLICATION OF THE ALGORITHM
(Raafat, 1991)
Increasing migration within our country accelerates the change In the pre-researches, the histograms of the undamaged and
in the usage of lands as well, especially in metropolitan areas. damaged buildings are analyzed. The typical forms of
To determine the changes, remote sensing techniques play an histograms are shown in Figure-1.
important role. This process requires using the data sets taking
in different dates. Environmental factors affect the reliable on
determination of changing.
One of the fundamental tasks in image understanding is the
automated extraction of man-made objects from aerial or space
borne imagery (Kim, 1999). The extraction of buildings in
particular, has attracted a lot of attention in photogrammetric
research (Lammni, 1999). A case in point is the photo- a
* Corresponding author. '
639 |