Detection of Geological Faults
Introduction
The geological faults are studied in geoscience, since are usually connected with seismic activity. In this work, we propose a framework for automatic enhancement and identification of geological fault structures which is faced for the first time. In experimental results, the proposed method has been tested on several offshore real topographic and bathymetric images yielding high performance results.
Methodology
1. The first module of the proposed scheme concerns the computation of Slope and Aspect images and their derivatives [1-2].
2. Next, we combine these images taking into account the fault properties, getting an enhancement image of the faults [1-2].
3. The faults are recognized by applying a modified version of the curvilinear structure detection algorithm [3].
Fig. 1: The first three steps of the method [2] and the schema of [1].
4. The enhancement and detection results of [1] are efficiently combined to construct a weighted graph that represents the possible fault points and the probability of belonging to the same fault [2].
5. In the last step, the most representative faults are provided by a suitable subset of the graph edges [2].
Fig. 2: An experimental result of [2]. Detected linear patterns of the faults on a region located in Heraklion region.
Downloads
- Poster of [1] in high resolution.
- Datasets used in [1-2] (.rar file). The datasets are saved in .mat files, see readme.txt for instructions.
Related Publications
[1] C. Panagiotakis and E. Kokinou, Automatic enhancement and detection of active sea faults from bathymetry,International Conference on Pattern Recognition, 2014 (accepted).
[2] C. Panagiotakis and E. Kokinou, Linear Pattern Detection of Geological Faults via a Topology and Shape Optimization Method, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014 (under revision).
[3] C. Panagiotakis, E. Kokinou and A. Sarris, Curvilinear Structure Enhancement and Detection in Geophysical Images, IEEE Trans. on Geoscience and Remote Sensing, vol. 49, no. 6, pp. 2040-2048, 2011.