Back to search

Aristotle University Of Thessaloniki (AUTH)

Combined use of geophysical data, satellite remote sensing data and geographic information systems (GIS) to locate and map archaeological relics

Abstract

dc:description

The research presented in this thesis is concerned with the problem of combining data from geophysical surveys and remote sensing for the detection and mapping of buried archaeological remains. The aim is to produce a reliable fusion method that presents the visual information, obtained from the geophysical data and the satellite images in a single fused image with higher information content. Αn effective registration method between two different geophysical images from different modalities is presented. The main characteristics of this method is that it performs local affine transformations, it utilizes the mutual information as a criterion of image similarity, it tapers the applied local affine transformation to ensure smooth transitioning, prevent discontinuities, and checkerboard artifacts and it is fast as it updates the values of subsets of pixels simultaneously, rather than one pixel at a time. Finally, although it uses a greedy optimization approach, it applies the deformations randomly, so it may be considered as a semi-stochastic method that, to some extent, it can avoid trapping in local maxima. The proposed method was developed using a pair of images, which are used to select the values of its parameters. It was then tested on totally different sets of images that refer to different potential archeological sites, corresponding to different time-periods and various soils –geological environments. Signal to noise ratio also varies significantly. In all cases, the images were pre-registered, using either their geographical coordinates, or a global rigid body transformation. The proposed method was then applied to correct for the local subtle image deformations. In all cases, the superposition of the registered images resulted to a sharper image than either of the two originals, with higher mutual information between the two images than that achieved by rigid body registration. Among the main objectives of this dissertation was the investigation of the fusion of dissimilar geophysical images produced from various geophysical methods. Although different images display the spatial distribution of different physical properties (ex. apparent resistivity, magnetic field gradient, electromagnetic reflections, etc.) they all can potentially detect and locate buried targets acting either additively or complementary to each other. Ideally, a fusion algorithm combines all the useful information of the initial images to one image, and suppresses the background noise. The resulted fused image improves the interpretation capabilities with respect to the initial images. We examined three different approaches of image fusion; the first simply averages the pixel values of the initial images. The second uses wavelet decomposition in 2-dimensions and then it selects the high energy coefficients to reconstruct one fused image. Finally, we developed a fusion algorithm based on the curvelet decomposition. The latter provides a various advantages in representing straight and curved features but more important, it allows the decomposition of an image in different bands (i.e. wavelengths), and also along different angles. Fusing the corresponding curvelet coefficients of the decomposed initial images results fused images with the maximum amount of information for each band and each orientation. Furthermore, in many archaeological applications of the geophysical methods the primary orientation of the potential targets is known (ex. from past excavations) or can be presumed (ex. historical evidence, geomorphology, etc.). Taking advantage of the curvelet decomposition, this orientation-information can be used as a-priori constraint to improve the results of fusion. We applied and compared the various fusion algorithms between geophysical images taken from different modalities and between geophysical images (electric resistivity measurements) and satellite images (Quickbird satellite image). Fusion using mean values performed decently almost in all cases combining satisfactory the information of the starting images. The problem with this method is that the resultants images appear smoother having lower contrast than the original ones. Noise is also propagates to the outcome image which is characterized from lower signal to noise ratio with respect to the initial images. In wavelet fusion transformations the images are separated into different sub-bands providing the capability to enhance the bands where the signal is expected or to smooth the ones with the noise. The resulted images had good contrast and sharpness compared with the results from the mean value fusion method. In some cases the images were very bright in certain areas (inherited from the initial images) causing loss of the information there. The advantage of curvelets compared to wavelets is that they make use of the ridgelet transform (using the Radon transform) to a partitioned version of the image so as the curved features to be represented as straight lines. This is the part where curvelets outperform to wavelets. The resulted images were clear with good contrast with high signal to noise ratio. In certain case targets that were invisible or difficult to distinguish in the initial images became apparent. Curvelet fusion method considering also the angle information gave the most effective resulted images especially in the case of linear features (straight lines-rectangular targets). Targets appear clearer and sharper.

Degree

thesis:*
Grantor dc:publisher
Aristotle University Of Thessaloniki (AUTH)
Year dc:date
2012

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Karamitrou, Alexandra
  • Καραμήτρου, Αλεξάνδρα

Subjects

dc:subject × 20

Rights

Language dc:language
eng

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/29840

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Karamitrou, Alexandra; Καραμήτρου, Αλεξάνδρα. Combined use of geophysical data, satellite remote sensing data and geographic information systems (GIS) to locate and map archaeological relics. Aristotle University Of Thessaloniki (AUTH), 2012. http://hdl.handle.net/10442/hedi/29840