Visualizing Aeromagnetic Texture Domains using Multichannel Filtering and Unsupervised Clustering
Abstract
Image texture is an intrinsic property of aeromagnetic data that together with the availability of closer survey line-spacings and more sensitive instrumentation, has become more valuable for qualitative interpretations. Problems arise though when it is used as a basis for geological interpretation, since visual textures are difficult to define. This can introduce bias when interpretations are not validated against repeatable and testable quantitative measures. We propose the use of well-established unsupervised texture segmentation techniques that cluster multi-scale, orientation-selective Gabor filters to automatically visualize different domains of textures in aeromagnetic maps. We outline how different methods of processing geophysical data can be integrated to create workflows that result in interpretable and repeatable models.
These workflows are applied to compiled aeromagnetic maps over the Palaeoproterozoic Earaheedy Basin, located over the northeastern margin of the Archean Yilgarn Craton and southeastern region of the Capricorn Orogen in Western Australia. Model selection and complexity using entropy is also analysed which highlight areas of uncertainty. These results are then validated against a geophysically constrained interpretation of the area and a review of the magnetic characteristics of the region. Our results demonstrate how automatic texture domaining from a single channel of aeromagnetic data. The methods and workflows are also made freely available as ArcGIS and MapInfo plugins in the Integrated Exploration Platform (https://www.waexplorationplatform.wa.edu.au/).- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2018
- Bibcode:
- 2018AGUFM.S11E0417N
- Keywords:
-
- 0555 Neural networks;
- fuzzy logic;
- machine learning;
- COMPUTATIONAL GEOPHYSICSDE: 1910 Data assimilation;
- integration and fusion;
- INFORMATICSDE: 1942 Machine learning;
- INFORMATICSDE: 7223 Earthquake interaction;
- forecasting;
- and prediction;
- SEISMOLOGY