HR: 1340h
AN: T43B-1345 [Abstracts]
TI: Application of the digital photogrammetry to obtain near-fault surface displacement field, eastern Taiwan
AU: * CHEN, R
EM: roufei@earth.sinica.edu.tw
AF: Rou-Fei Chen, Institute of Earth Sciences, Academia Sinica, P.O. box1-55, Nankang,
Taipei, 115, Taiwan
AU: LEE, J
AU: CHANG, K
AU: CHAN, Y
AB:
Measuring earthquake surface deformation provides information of both general
kinematics of earthquake faulting and secondary deformation features, which usually
reflect fault behaviors and surface geological heterogeneity. Recently, technology
provided increasing variety of methods to measure surface deformation of
earthquakes. For instance, photogrammetic analysis, especially sub-pixel correlation
of remote sensing imagery, which we focused on in this study, is useful to reconstruct
the co-seismic surface displacement field. Photogrammetry represents the process of
deriving topographic information about an object through measurements made on
aerial images of the object.
Aerial photographs acquired using a high-resolution digital images and ground survey has been to monitor
changes in topographic features and surface deformation, and to obtain the digital elevation models (DEMs). To
rectify the image data, it is necessary to obtain geometrical parameters including camera type, frame focus, and
aerial height. Furthermore, it is important to achieve a reasonable accuracy without a great number of ground
control points (GCPs) when dealing with large-scale imagery. In order to obtain better ground controls for
geometric calibration, we carried out RTK-GPS measurements at each ground points with a resolution of a few
centimeters in horizontal and vertical components. Therefore, GCPs provide strong geometrical constraints for
triangulation and orthorectification of images. This measurement is crucial because otherwise it is almost
impossible to obtain orthorectified photographs with a precision in pixel level as good as 10-20 cm. The particle
image velocimetry (PIV) method is an application of an optical image (e.g., aerial photographs and SPOT images)
correlation technique, in which the horizontal displacement field is measured by comparison of images acquired
at two different times. This method is based on a sub-pixel correlation of orthorectified images, using sliding
windows. Each image is divided into several small areas (windows). The cross-correlation function is a pattern-
matching routine that determines the relative displacement between images, which are shifted according to the
best overlap. The residual offsets of image pairs are computed from the phase shift of the Fast Fourier Transform
of the sliding window. This value corresponds to the horizontal displacement vector of the window. The
correlation between two images can be efficiently performed as long as their texture is similar. The creation of
DEMs from oblique and non-metric imagery using automated digital photogrammetry can be difficult. Recently,
technology provided increasing variety of methods, especially seamless DEMs are generated for entire blocks or
for any sub-block or polygon area. In this study, we focused to put forward a survey method to monitoring
morphological change through reconstruct the high precision DEMs and measure surface displacement.
DE: 6924 Interferometry (1207, 1209, 1242)
DE: 8175 Tectonics and landscape evolution
SC: Tectonophysics [T]
MN: 2007 Fall Meeting