![Hier klicken, um den Treffer aus der Auswahl zu entfernen](images/unchecked.gif) |
Titel |
Topographic attributes as a guide for automated detection or highlighting of geological features |
VerfasserIn |
Sophie Viseur, Thibaud Le Men, Yves Guglielmi |
Konferenz |
EGU General Assembly 2015
|
Medientyp |
Artikel
|
Sprache |
Englisch
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 17 (2015) |
Datensatznummer |
250109123
|
Publikation (Nr.) |
EGU/EGU2015-11583.pdf |
|
|
|
Zusammenfassung |
Photogrammetry or LIDAR technology combined with photography allow geoscientists to
obtain 3D high-resolution numerical representations of outcrops, generally termed as Digital
Outcrop Models (DOM). For over a decade, these 3D numerical outcrops serve as support for
precise and accurate interpretations of geological features such as fracture traces or plans,
strata, facies mapping, etc. These interpretations have the benefit to be directly
georeferenced and embedded into the 3D space. They are then easily integrated
into GIS or geomodeler softwares for modelling in 3D the subsurface geological
structures.
However, numerical outcrops generally represent huge data sets that are heavy to
manipulate and hence to interpret. This may be particularly tedious as soon as several scales
of geological features must be investigated or as geological features are very dense and
imbricated. Automated tools for interpreting geological features from DOMs would be then a
significant help to process these kinds of data. Such technologies are commonly used for
interpreting seismic or medical data. However, it may be noticed that even if many efforts
have been devoted to easily and accurately acquire 3D topographic point clouds and photos
and to visualize accurate 3D textured DOMs, few attentions have been paid to the
development of algorithms for automated detection of the geological structures from
DOMs.
The automatic detection of objects on numerical data generally assumes that signals or
attributes computed from this data allows the recognition of the targeted object boundaries.
The first step consists then in defining attributes that highlight the objects or their
boundaries. For DOM interpretations, some authors proposed to use differential operators
computed on the surface such as normal or curvatures. These methods generally
extract polylines corresponding to fracture traces or bed limits. Other approaches rely
on the PCA technology to segregate different topographic plans. This approach
assume that structural or sedimentary features coincide with topographic surface
parts.
In this work, several topographic attributes are proposed to highlight geological features
on outcrops. Among them, differential operators are used but also combined and processed to
display particular topographic shapes. Moreover, two kinds of attributes are used:
unsupervised and supervised attributes. The supervised attributes integrate an a
priori knowledge about the objects to extract (e.g.: a preferential orientation of
fracture surfaces, etc.). This strategy may be compared to the one used for seismic
interpretation. Indeed, many seismic attributes have been proposed to highlight
geological structures hardly observable due to data noise. The same issue exist with
topographic data: plants, erosions, etc. generate noise that make interpretation sometimes
hard.
The proposed approach has been applied on real case studies to show how it could help
the interpretation of geological features. The obtained “topographic attributes” are shown and
discussed. |
|
|
|
|
|