- Email: [email protected]
Geo3DML: A standard-based exchange format for 3D geological models
Accepted Manuscript Geo3DML: A standard-based exchange format for 3D geological models Zhangang Wang, Honggang Qu, Zixing Wu, Xianghong Wang PII:
S0098-3004(17)30271-6
DOI:
10.1016/j.cageo.2017.09.008
Reference:
CAGEO 4022
To appear in:
Computers and Geosciences
Received Date: 8 March 2017 Revised Date:
25 August 2017
Accepted Date: 15 September 2017
Please cite this article as: Wang, Z., Qu, H., Wu, Z., Wang, X., Geo3DML: A standard-based exchange format for 3D geological models, Computers and Geosciences (2017), doi: 10.1016/ j.cageo.2017.09.008. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1 2 3
5 6 7 8
Geo3DML: A standard-based exchange format for 3D geological models
RI PT
4
Zhangang Wang a*, Honggang Qub, Zixing Wub, Xianghong Wangb a
College of Geosciences and Surveying Engineering, China University of Mining and Technology, Beijing, 100083 China b
Development & Research Centre, China Geological Survey, Beijing, 100037, China *
(Corresponding author: [email protected])
SC
9
Abstract: A geological model (geomodel) in three-dimensional (3D) space is a digital representation
11
of the Earth’s subsurface, recognized by geologists and stored in resultant geological data (geodata).
12
The
13
addressed by developing standard-based exchange formats for the representation of not only a single
14
geological object, but also holistic geomodels. However, current standards such as GeoSciML cannot
15
incorporate all the geomodel-related information. This paper presents Geo3DML for the exchange of
16
3D geomodels based on the existing Open Geospatial Consortium (OGC) standards. Geo3DML is
17
based on a unified and formal representation of structural models, attribute models and hierarchical
18
structures of interpreted resultant geodata in different dimensional views, including drills,
19
cross-sections/geomaps and 3D models, which is compatible with the conceptual model of GeoSciML.
20
Geo3DML aims to encode all geomodel-related information integrally in one framework, including
21
the semantic and geometric information of geoobjects and their relationships, as well as visual
22
information. At present, Geo3DML and some supporting tools have been released as a data-exchange
23
standard by the China Geological Survey (CGS).
EP
TE D
increasing demand for data management and interoperable applications of geomodelscan be
AC C
24 25 26
M AN U
10
Keywords: Geo3DML; geological model; standard; data exchange
27 28
1. Introduction
29
3D geological modeling methods have matured in recent years and have been widely applied (Mallet, 1997; De
30
Kemp, 1999; Lemon and Jones, 2003; Sprague and De Kemp, 2005; Apel, 2006; Caumon et al., 2009; Caumon,
ACCEPTED MANUSCRIPT 2010; Collon et al., 2015). Geological survey organizations (GSOs) from different countries have put forward
32
3D geological modeling and mapping programs, which aim toward building a 3D geological framework that
33
provides a basic platform for full 3D cognition of the subsurface (DGSM, 2005; USGS, 2007; Kessler et al.,
34
2009; Berg et al., 2011; CSIRO, 2012; Gupta et al., 2015). In general, from the view of geological recognition
35
or data observation and collection, so-called geological data (geodata) mainly include boreholes, cross-sections,
36
geologic maps and 3D models and these data are actually specific forms of geological models (geomodels),
37
which provide an interpretative cognition and digital representation of Earth’s subsurface (Mallet, 2002). The
38
continuing progress in the generation and analytical methods of geomodels or geodata creates strong
39
expectations on agencies, especially geological survey organizations (GSOs), for exchanging and sharing
40
re-usable geodata, just as OneGeology (2008) shares 2D geoscience data via the internet.
SC
RI PT
31
It is necessary to provide standardized information encoding rules, making data providers and users process
42
the contents of geodata in a consistent way (Howard, 2009). The contents of a single geological object
43
(geoobject) generally include two important and distinct aspects: (1) human words or vocabulary-based
44
conceptual attributes; and (2) location or shape related geometric information. Currently, there are several
45
standards developed to describe the geoobject-related contents. GeoSciML (Sen and Duffy, 2005) defined very
46
basic geological data structures and document formats based on GML (2007), and GeoSciML 4.1 has now been
47
adopted as an OGC standard (GeoSciML, 2016). The geometry data type from GML can be used to represent
48
the shape of a geoobject. RESQML (2012) is mainly used for data exchange of 3D reservoir objects in the oil
49
and gas exploration industry. However, these still lack holistic geomodel representations. The purpose of
50
designing an exchange standard is more than providing a readable file format. The holistic representation of
51
geomodels encompasses the rules of space partitioning, the semantic structures of geoobjects and the
52
identification of relations between geoobjects (Wang et al., 2016), as well as other information for exchange,
53
including metadata, quality evaluation, visual information, etc. All such information can be considered as
54
geomodel-related data for reuse, and should be organized and encoded in a unified framework. Thus, Geo3DML
55
is proposed to provide a standardized representation and exchange format for 3D geomodels based on the
56
existing standards.
AC C
EP
TE D
M AN U
41
57
The remainder of this paper is structured as follows. The design principles of Geo3DML are given in Section
58
2. Section 3 introduces a unified and formal method of representing interpreted resultant geomodels in different
59
dimensional views. Section 4 introduces the framework and main components of Geo3DML. Section 5 gives
ACCEPTED MANUSCRIPT 60
information relating to supporting tools. Section 6 discusses the features and limits of Geo3DML. Section 7
61
presents our conclusions and plans for future work.
62
2. Design principles
63 Geological survey organizations
Users
Institutes & Enterprises
Software1 WVS
3D Geological Spatial Database
PDF3D
Data sharing
Government Departments
……
Other Consumers
M AN U
……
WFS
SC
Spatial Database
Geo3DML
Integration
Software3
Geo3DML
3D Geological Software2
RI PT
3D Geological Modeling and Analysis Tools
Exchange
64 65 66 67 68 69 70 71 72 73 74 75 76 77
Management platform
……
Fig. 1. Requirements for exchanging and sharing geomodels using Geo3DML for GSOs
One of the main responsibilities of GSOs is to manage processed digital geodata and to promote geodata
79
services in different applications, as shown in Fig. 1(dash line). The geodata can not only realize the conceptual
80
description of subsurface information, but can also record visual information. Geo3DML
81
encoding of geomodels in 3D geospace, is designed to transfer geodata in different stages, encompassing: (i)
82
exchanging geomodels between different software; (ii) ensuring geomodels integrated within spatial databases;
83
and (iii) sharing geomodels in a standardized service-oriented architecture, such as WFS (2005) and WVS
84
(2010), or other applications, e.g., PDF3D (2007). Then, users can download, visualize or query the geodata.
85
Thus, the principles in creating Geo3DML include:
EP
87
Platform independent and integrated data model. As a public exchange format, Geo3DML is
independent of specific geomodeling software. Moreover, Geo3DML provides a unified way to express
88
structural models and attribute models and organize the hierarchical structure of geodata, including
89 90
through standard
AC C
86
TE D
78
drills, cross-sections/geomaps and 3D models (see Section 3).
Segregation of geomodel data and visual information. The core framework of Geo3DML in Section 4.1
91
is based on a loosely coupled segregation policy, such that Geo3DML can support modular data
92
exchange. In other words, geodata can be separated or shared as needed. For example, data users can
93
obtain a complete geomodel without visual information.
ACCEPTED MANUSCRIPT 94
Reusing and extending existing standards. Geo3DML mainly uses the existing OGC standards (Table
95
1), based on GML extensions, and the conceptual model of geoobjects is derived from OGC
96
GeoSciML. Thus, Geo3DML is allowed compatibility with the existing OGC web-service
97
architectures.
Standard name Filter Encoding
Version 1.1.0
GeoSciML
Geoscience Markup Language
4.1
GMD
Geographic Information-Metadata
GML
Geography Markup Language
3.2.1
SE
Symbology Encoding
1.1.0
SWE
Sensor Web Enablement Common Data Model
2.0
100 101
Contribution in Geo3DML Extension for encoding 3D visual parameters Definition of geoobjects and their relations Extension for encoding project and geomodel metadata Extension for encoding geometry-related shape objects and field data Extension for encoding 3D visual parameters Definition of the attribute structure of geoobjects
SC
Abbreviation FE
RI PT
Table 1. Relevant standards used in Geo3DML
M AN U
98 99
Modularity and extendibility. Geo3DML uses the XML Schema language and UML to define the
102
framework (see Section 4.1), which mainly includes the seven modules shown in Fig. 2. The function
103
of each module is given in Table 2.
<>
Geo3DProject
TE D
GeoMetadata
<>
<>
<>
104 105 106 107
AC C
EP
GeoGeometry
GeoFeature
Geo3DStyle
<>
GeoBasicType
<>
GeoProperty
Fig.2. Modular components of Geo3DML
Table 2. Description of the Geo3DML modules No 1
Module name Geo3DProject
2
GeoMetadata
3
Geo3DStyle
4
GeoFeature
5
GeoGeometry
Description Defines the project information, and the structure of geomodels. This is the main component of the data transfer format defined by Geo3DML Defines the structure of the GeoModel metadata Defines the 3D visualization parameters and the structure of the parameter library Defines the geological features and the structure of the classes of geological objects Defines the extended structure of geometric objects stated in the GML specifications
ACCEPTED MANUSCRIPT
108 109
6
GeoProperty
Defines the extended structure stated in gmlcov:Abstract-Coverage and is used to describe the 3D property fields
7
GeoBasicType
Defines the structure of the basic data used by Geo3DML
Ease of use. A series of support tools have been developed in order to help users to check the veracity of Geo3DML files, visualize geomodels and program Read-Write interfaces.
3. Methods of representing geomodels
111
3.1 Three levels for representing geomodels
112
The formal representation of geomodels is the foundation of Geo3DML, providing a unified framework to
113
integrate geological information. Several studies attempted to represent geomodels, but their focus was on
114
constructing and editing a geological model, such as the G-maps (Lienhardt, 1994; Mallet, 2002), the Sealed
115
Geological Model (Caumon et al., 2004) and the Wire frame (Xu and Tian, 2009). The geomodels represented
116
by G-maps have complete mathematical theories to describe spatial data in any spatial dimensions, which have
117
been extended into the spatio-temporal domain (Le et al., 2013). More relevant discussions about these models
118
were given in Wu and Caumon's studies (Wu, 2004; Caumon, 2010). Mallet (2002) noted that a geomodel
119
provides an abstract digital representation of the Earth’s subsurface, composed of both a geological structural
120
model, which stores the shapes and relations of geoobjects utilizing surface-based boundary representations, and
121
an associated attribute model, used to describe the material properties of these geoobjects. Such representation
122
involves spatial partitioning under specific conditions by constructing a mapping relationship between the
123
segmented spaces and geoobjects (Wang et al., 2016). Geomodels consist of a set of exclusive and
124
interconnected geoobjects, which define their geological semantics and geometric properties, and specify
125
topological relationships and self-consistencies. Moreover, a geomodel is recognized and processed by
126
geologists, and appears in interpreted resultant geodata, such as boreholes, sections/geomaps, and 3D models
127
(Fig. 3), which are conventionally portrayed on maps or rendered in a 3D scene. These data can be used to
128
understand geological phenomenon along a line path, a surface, or in 3D space. Thus, geomodels provide
129
geological cognition of the subsurface in various dimensional views and have three levels (Fig. 4): (i) spatial
130
modeling is for modeling the geometric shapes of geoobjects, and uses vector or raster data types to portray
131
locations and graphic characteristics; (ii) object modeling is for modeling single, spatially referenced geoobjects,
132
where an object models the information about itself, particularly a representation of the geological setting and
133
concepts, essential attributes and linked geometries as spatial properties; and (iii) integrated modeling is for
134
modeling spatially related collections and repositories of geoobjects and for forming geomodels with explicit
AC C
EP
TE D
M AN U
SC
RI PT
110
ACCEPTED MANUSCRIPT spatial relations between objects and model-related information.
136 137 138 139 140 141 142 143 144 145 146 147 148 149
3D
Vertex Interface
Edge Domain
SC
Contact
East
Dimension of Geometric elements
of GeoModel
3
3 (3D model)
Body
2 (map, section) 1 (drill, traverse) The ‘
2
1
0
Interface
Edge
Vertex
Domain
Boundary
Junction
Interval
Contact
’ arrow means the ‘equivalent’ representation of elements
Object modeling
Spatially related collections (geomodel)
EP
Integrated modeling
TE D
Fig. 3. Geometric elements of geomodels in one, two and three dimensional views (Burns, 1988 and Thiele et al., 2016)
Spatially-referenced geological objects Id,name properties
AC C
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169
RI PT
Body
Dimension
150 151 152 153
1D
2D
M AN U
135
Spatial modeling
Stratum (Jurassic)
Geometric objects
Model types for interpreted resultant geological data: borehole, cross-section, geomap, 3D model ... Purpose: organizing and integrating the geomodel-related information in a unified way.
Object types for geological concepts: unit, fault, contact …
geologic
Purpose: defining data structures and vocabularies using GeoSciML
Geometry types for Boundaryrepresentation and Gridrepresentation: point, line, triangle (polygon), polyhedron ... Purpose: partitioning geopace and modeling geometric shapes of geoobjects by extending GML
Fig. 4. Three levels of geomodels in Geo3DML
170
A sample of a 3D model, composed of both a structural model and associated attribute model, is shown in Fig.
171
5. The modeling process involves the partitioning and discretization of the 3D subsurface. The boundary
ACCEPTED MANUSCRIPT surfaces of the structural model are generated based on a number of geological boundary conditions (e.g., faults
173
or horizons), and then each complete geoobject is closely formed by the corresponding surfaces. In current
174
practice, the representation of geoobjects can differ in their semantic and spatial relationships. For example (Fig.
175
5c, d and e), three possible common types of topological representations of geo-unit A can be categorized, where
176
the third method represents explicit topological relationships between A and geological boundaries. The attribute
177
model of this sample is constructed using a grid/voxel-based model, supporting discretization of the space of a
178
sealed geo-unit A (Fig. 5f and g) to generate the distribution of geological attribute fields (Caers et al., 1999;
179
Mallet, 2002; Caumon, 2010).
RI PT
172
M AN U
SC
3D Model
Stratum A
(a)
180 181
(b)
182 (c)
(d)
(e)
Grid/Voxel-based representation
AC C
183 184
EP
TE D
Surface-based representation
185 186 187 188 189 190
(f)
(g)
Fig. 5. Methods of representation of a 3D geomodel: (a) A 3D structural geomodel; (b) Wireframe representation of stratum A using triangulated irregular networks (TINs); (c) A is formed using one fully-closed surface; (d) A is composed of two components split by a fault; (e) A is topologically formed by combining boundary surfaces; (f) A discrete attribute model
ACCEPTED MANUSCRIPT embedded in stratum A using tetrahedral meshes; and (g) An attribute model using cornerpoint grids.
193
3.2 Integrated model
194
A geomodel is first and foremost a spatial data model. A full description of a spatial model involves three
195
aspects: spatial partitioning, constructive rules, and supported objects and primitives (Zlatanova et al., 2004). As
196
introduced in our work for 3D models (Wang et al., 2016), geomodels in n-dimensional views under the same
197
time and observation scales can be formally given:
198
n-GeoModel = {Gp, O, R, A} (n = 0, 1, 2 and 3)
RI PT
191 192
(1)
Gp represents geometric discretization and partitioning of subsurface space.
200
O represents geoobjects (i.e., GeologicFeatures in GeoSciML and Geo3DML). Wang et al. (2016) gave
SC
199
the constructive rules and formal definitions of the basic types of geoobjects—GeologicBoundary,
202
GeologicUnit and GeologicStructure. GeologicBoundary is a special type of GeologicStructure. It is an
203
abstract representation of all geological boundary constraints for spatial partitioning, such as a part of a
204
fault or a horizon (stratigraphic/lithological boundary) in an actual model. For a GeologicUnit object o,
205
the geometrical dimension dim(o) = n, and for a GeologicStructure, dim(o) = n-1.
R represents explicit spatial relationships between geoobjects. The formal representation of spatial
TE D
206
M AN U
201
relations is becoming increasingly important for alphanumeric query and analysis of spatial data
208
(Bradley and Paul, 2014; Leopold et al., 2015; Thiele et al., 2016). GeoSciML has defined several
209
abstract geological relations and spatial relations. Geo3DML emphasizes the topological relations and
210
directional relations, where R = {above, below, at, boundary, composition}, as defined in Geoscience
211
Spatial Framework (GSF) (DGSM, 2005) from BGS and GSIS (Ming et al., 2010). The ‘above’ or
212
‘below’ specifies a GeologicBoundary adjacent to a unique GeologicUnit above or below it. The ‘at’
AC C
EP
207
213
denotes a GeologicBoundary interpreted as a more specific GeologicStructure. The ‘boundary’
214
specifies a relationship of the bounds of a GeologicUnit composed of GeologicBoundaries. The
215
‘composition’ specifies a relationship of a GeologiUnit composed of other GeologicUnits, or the
216 217
relationship of a GeologicStructure composed of other GeologicStructures.
A represents geological attribute fields (v = F(x)) discretized and bound to a geometry object G using a
218
grid/voxel-based representation. Any scalar, vector or tensor field can be discretely stored in a primitive
219
(minimum cell) of G through four binding types as shown in Fig. 6.
220
ACCEPTED MANUSCRIPT
221 Attributes on:
Vertex
Edge
Face
Cell
RI PT
Fig. 6. Four binding types for storing the value of attribute fields in a geometric cell
Generally, n-GeoModels can be spatially represented in a unified form (Table 3) and the geological structures and units are presented by n-1- and n-Geometries, respectively. The n-Geometries can be further discretized into n-primitives, or other compatible Cartesian grids, to form the attribute models.Table 3. Spatial representations of an n-GeoModel in various dimensions n-GeoModel
Shape of GeologicBoundary
Shape of GeologicUnit
Attribute model
SC
n
(n-1-Geometry)
(n-Geometry)
(Primitives embedded in n- Geometry) Node
0
Observed Point
/
Point
1
Borehole
Point
Line
2
Cross-section/Geomap
Line
Polygon
Grid (Triangle)
3
3D Model
Surface
Body
Polyhedron (Grid3D)
M AN U
222 223 224 225 226 227 228
Segment
4. Components of Geo3DML
230
4.1 Framework
231
The top-level components of Geo3DML are shown in Fig. 7, and the core framework is composed of GeoModel
232
and Geo3DMap. The class GeoModel is designed to represent the n-GeoModels (eq. (1)). The class Geo3DMap
233
is used to define how to visualize the geomodels. Geomodels are usually created in a geological survey
234
campaign or project, and the class Geo3DProject is used to organize all the relevant GeoModels and
235
Geo3DMaps. Geo3DProject also records the project information such as location, extent and responsible
236
parties.
AC C
EP
TE D
229
237
The configuration of Geo3DMap draws on the concept of the traditional map in 2D geographic information
238
systems (GISs) and is composed of several layers. Geo3DLayer is a layer that is used for the combination of a
239
set of geoobjects (GeoFeatureClass) and their visual information (Geo3DStyle). A GeoFeatureClass is capable
240
of being visualized by a variety of Geo3DStyles, such as color or texture filling styles.
241
Geo3DML employs the fundamental class GeologicFeature from GeoSciML to model all geoobjects in a
242
geomodel. However, the data structures (semantic attributes) of GeologicFeature defined by GeoSciML are
243
fixed. Although GeoSciML-Lite (GeoSciML, 2016) provides an extendible way to allow additional attributes, it
244
is still insufficient to exchange generic feature types. Geo3DML provides another choice for defining
ACCEPTED MANUSCRIPT 245
GeoFeatures (see Section 4.2). GeoFeatureClass is a set of GeologicFeatures with the same attribute
246
configuration.
247
GeologicFeatureRelation from GeoSciML, as shown in Fig. 8. The model-related metadata, such as describing
248
the project initiative and data sources or the process of constructing the geomodels, are given by MD_Metadata
249
from GMD (2007).
five
spatial
relationships
mentioned
in
Section
250 1
Geo3DProject (f rom Geo3DML)
GeoModel
M AN U
(f rom Geo3DML)
+FeatureClasses
GeoFeatureClass (f rom Geo3DML)
1
+FeatureClass
1
choice
+Features 0..n GeoFeature
(f rom GeoSciML)
(f rom Geo3DML)
TE D
+Features 0..n GeologicFeature
+Layers
EP
1
0..n
Geo3DLayer (f rom Geo3DML)
1 1
+Styles
0..n
Geo3DStyle (f rom Geo3DML)
Fig. 7. UML class diagram: framework of top-level components
AC C
251 252 253
1
0..n
Geo3DMap
(f rom Geo3DML)
0..n
derived
+Maps
+Models
1
are
SC
0..n
1
3.2
from the
RI PT
The
ACCEPTED MANUSCRIPT GeoModelMetadata
MD_Metadata
(from Geo3DML)
(from GM D)
GeologicFeatureRelation
GeoModel
(from GeoSciML)
(f rom Geo3DML)
0..n
0..n
Geo3DProject
1
1
Geo3DMap
0..n
(f rom Geo3DML)
(f rom Geo3DML)
1 1
RI PT
1
+Layers
+FeatureClasses
0..n
1
GeoFeatureClass
1
0..n
Geo3DLayer
(f rom Geo3DML)
SC
(f rom Geo3DML)
BoundaryRelationship
DefiningStructure
BelowRelationship
AboveRelationship
CompositionRelationship
(from GeoSciML)
(from GeoSciML)
(from Geo3DML)
(from Geo3DML)
(from Geo3DML)
254 255 256
M AN U
Fig. 8. UML class diagram: metadata and relationships
4.2 Semantic and geometric attributes of geologic features
258
The procedure for modeling a single generic geoobject (Fig. 4) involves the definitions of semantic attributes,
259
geometry data and attribute fields. Fig. 9 shows the data model of geologic features.
Feature choice: GeoFeature or subclass of GeologicFeatrue from GeoSciML
0..n
TE D
257
AbstractFeature
AbstractGeometry
AbstractDiscreteCoverage
(from GM L)
(from GML)
(from GMLCov)
GeologicFeature (f rom GeoSciML)
1
0..n
1 +Shape 1
MappedFeature (from GeoSciML)
EP
+occurrence
+Features 1 GeoFeatureClass (from Geo3DML)
GeoDiscreteCoverage
AC C
GeoFeature
Schema Features FeatureType SpatialType Dimension
1
0..n
+Features
(from Geo3DML)
(from Geo3DML)
1
Fields Geometry
0..1
Geometry (from Geo3DML)
1
0..n
+ShapeProperty
PropertyFieldType : CodeList SamplingFrame : AbstractGeometry SamplingTarget : CodeList
1
1
0..1 +Schema DataRecord (from SWE)
field
260 261 262 263 264
0..n
1 1..n +field
+Fields
AbstractDataComponent (from SWE)
Coldelist of SamplingT arget VERTEX EDGE FACE CELL Coldelist of PropertyFieldT ype SCALAR VECT OR TENSOR MAT RIX
Fig. 9. UML class diagram: representation of geologic features
ACCEPTED MANUSCRIPT AbstractGeometry (from GML)
AbstractGeometricAggregate
AbstractGeometricPrimitive
Grid
(from GML)
(from GM L)
(from GML)
1
GeoGrid
1
(from Geo3DML)
+grid
1 1 +transformationMatrix AbstractCurve
AbstractSurface
DirectPosition
AbstractSolid
TransformationMatrix4x4
(from GML)
(from GM L)
(from GM L)
(from GML)
(from Geo3DML)
GeoTin
(from GM L)
(from Geo3DML)
Vertices Triangles
1
0..n
+Vertices
Vertex (from Geo3DML)
IndexNo
0..n
1
GeoVolume Vertices
+Vertices
1 GeoTetrahedronVolume (from Geo3DML)
Tetrahedrons
GeoCuboidVolume (from Geo3DML)
M AN U
Cuboids
1
0..n
+Triangles
Triangle IndexNo VertexList NeighborList
265
1
GeoCornerPointGrid (from Geo3DML)
Dimension Pillars Cells
0..n
+Cuboids
(from Geo3DML)
Cuboid
IndexNo VertexList NeighborList
(from Geo3DML)
IndexNo VertexList
TE D
Fig. 10. UML class diagram: geometry
267 268
(from GM L)
+Tetrahedrons
Tetrahedron
(from Geo3DML)
266
0..n
Solid
(from Geo3DML)
SC
LineString
RI PT
Point (from GM L)
•
There are two choices for extending attributes of a GeologicFeature in Geo3DML. One is to use the specific GeoSciML subclass of GeologicFeature. The attribute FeatureType of GeoFeatureClass
270
indicates the specific class name, e.g., GeologicUnit or Contact. Another makes use of the GeoFeature
271
from Geo3DML to add user-defined attributes. The attribute Schema of GeoFeatureClass is an instance
272
of swe:DataRecord from OGC SWE specifications (SWE, 2009), and is used for defining the structures
273
of the attribute table. The attribute Fields of GeoFeature record the value of each corresponding field
274 275
AC C
EP
269
in Schema.
•
The geometry-related shape objects and field data extend from the GML geometry and coverage
276
definitions (GML, 2007). The shape of a GeologicFeature object, i.e., MappedFeature, is associated
277
with an instance of gml:AbstractGeometry. GeoSciML only focuses on 2D geometry portrayed in
278
geologic maps. However, MappedFeature cannot support coverages and hardly covers well-known
279
geometric types used for surface-based and grid-based representations in 3D geomodeling software.
280
Geo3DML defines the Geometry class to extend MappedFeature and provides more geometric types to
ACCEPTED MANUSCRIPT 281
represent
282
tetrahedron-based model, cuboid-based model and cornerpoint grid, as shown in Fig. 10. Each
283
geometry object can be stored in binary form in XML-based documents using OGC Well-Known
284
Binary (WKB) encoding.
including
structured
and
unstructured
meshes,
such
as
the
Field data are dependent on a specific geometry object. Geo3DML defines GeoDiscreteCoverage,
RI PT
•
285
GeologicFeatures,
which inherits from gmlcov:AbstractDiscreteCoverage, to support the geological attribute model. It has
287
three main extension attributes: (1) SamplingFrame points to an associated geometry object; (2)
288
SamplingTarget refers to one of four binding types depicted in Fig. 6; and (3) PropertyFieldType refers
289
to the field type, such as scalar, vector or tensor.
SC
286
290 4.3 Geo3DStyle
292
The visual information of a geomodel is recorded in Geo3DStyles. The core model is illustrated in Fig. 11. By
293
referring to SLD (2005), CityGML (2012) and X3D (2004), the extension for encoding 3D visual parameters
294
comes from the SE (2006) and FilterEncoding (FE, 2005) specifications. FeatureTypeStyle is used to define the
295
visual parameters of geometric data, and CoverageStyle is used to define the visual parameters of the attribute
296
fields. Rule classifies the specific visual parameters (Symbolizer) based on different geological attributes using
297
Filter and ElseFilter. To support 3D visualization, Geo3DML gives common encoding of 3D visual information
298
for
299
(geo3dml:GeoSurfaceSymbolizer), and field data (geo3dml:GeoDiscreteCoverage).
(geo3dml:GeoPointSymbolizer),
AC C
EP
point
TE D
M AN U
291
line
(geo3dml:GeoLineSymbolizer),
surface
ACCEPTED MANUSCRIPT AbstractGML (from GML)
0..1
1
Geo3DStyle
(from SE)
1
CoverageStyle
0..1
(from Geo3DML)
1
(from SE)
1 0..n
0..n Rule (from SE)
Filter (from Fi lterEncodi ng)
0..1
1
1 1
ElseFilter (from SE)
0..1
1..n Symbolizer
LineSymbolizer
(from SE)
(from SE)
GeoPointSymbolizer
GeoLineSymbolizer
(from Geo3DML)
(from Geo3DML)
1
1
GeoDiscreteCoverageSymbolizer
(from Geo3DML)
(from Geo3DML)
1
1
0..2
0..2
Material
AbstractTexture
GeoSceneStyle
(from Geo3DML)
(from Geo3DML)
EP AC C
301
GeoSurfaceSymbolizer
(from Geo3DML)
AmbientIntensity DiffuseColor EmissiveColor Shininess SpecularColor Transparency
300
(from SE)
TE D
1
1
RasterSymbolizer
M AN U
PointSymbolizer
SC
(from SE)
RI PT
FeatureTypeStyle
GeoReferencedTexture
ParameterizedTexture
(from Geo3DML)
(from Geo3DML)
Fig. 11. UML class diagram: Geo3DStyle
302
5. Supporting tools
303
Geo3DML (Version 1.0) was accepted and released in December 2015 by CGS as a standardized data-exchange
304
format for 3D geomodels. It is intended that Geo3DML will ultimately encompass a broad range of
305
requirements for geomodel data interoperability. In the process of compiling the Geo3DML, five geomodeling
306
software companies in China were invited to participate in the validation from multiple geological domains,
307
including fundamental geology, hydrogeology, urban geology, mineral geology and so on. In order to facilitate
308
the easy development of Geo3DML by different users, a series of supporting tools have also been provided,
ACCEPTED MANUSCRIPT including: (1) Geo3DML Viewer is a tool to check the veracity of Geo3DML files and to visualize the data
310
running on Windows; (2) Geo3DML SDK is an open-source C++/C# program for developing Read-Write
311
interfaces for Geo3DML-based documents; and (3) other tools for data exchange between Geo3DML and other
312
well-known file formats, such as the PDF3D convertor and GoCAD convertor. As shown in Fig. 12, for
313
example, the 3D bedrock geomodel of Beijing generated by GSIS (Ming et al., 2010) was encoded using
314
Geo3DML and rendered by Geo3DML Viewer and GoCAD (Version 2011, now named SKUA), respectively.
315
The differences in visual effect between the two screenshots are due to the different settings of global scene
316
rendering parameters, such as the number of light sources and ambient light. These parameters
317
exchanged in Geo3DML.
SC
RI PT
309
are not
In addition, a protosystem was developed to visualize Geo3DML-based geomodels in a web-service
319
application implemented using WebGL (2014). All such information can be obtained from the Geo3DML
320
website www.geo3dml.cn/en.
321 322
(a)
AC C
EP
TE D
M AN U
318
RI PT
ACCEPTED MANUSCRIPT
SC
323 (b)
325
Fig. 12. The 3D bedrock geomodel of Beijing generated by GSIS (Ming et al., 2010), and rendered in: (a) Geo3DML
326
Viewer; and (b) GoCAD.
327
6. Discussion
328
At present, Geo3DML focuses on the exchange of geomodels—3D interpreted resultant geodata, serving GSOs.
329
The purpose is to provide a standardized representation and encoding method based on the existing standards,
330
especially representing geomodels associated with the conceptual model of GeoSciML. Geo3DML considers
331
geomodels integrally to organize all the information in a unified way, as well as visual information. This feature
332
is the main difference with other standards.
TE D
M AN U
324
GeoSciML is most suited for geoobject modeling (Fig. 4), supporting descriptions of geological setting and
334
concepts. However, the geometry of a GeologicFeature cannot be well modeled by the popular methods of
335
boundary-based and grid-based representations. The current extensions of GeoSciML are always implemented
336
to describe the conceptual attributes of geoobjects in specific geological domains, such as GWL (Boisvert and
337
Brodaric, 2007) for hydrology, INSIPRE (2008) for multiple disciplines, GeoSciGraph (Gupta et al., 2015) and
338
3D-GEM (Tegtmeier et al., 2014) for geotechnics. The above researches hardly mention 3D geometry related
339
extensions.
AC C
EP
333
340
RESQML is used for the data exchange of 3D reservoir objects (RESQML, 2012). The main feature of
341
RESQML is the relatively complete definitions of various types of 3D grids for numerical reservoir simulation.
342
RESQML only defines some common geoobjects, like horizon and fault, which are derived from geophysical
ACCEPTED MANUSCRIPT 343
data interpretation. So the data structures of geoobjects are different from that of GeoSciML. Meanwhile, their
344
definitions do not use or extend OGC GML or GeoSciML.
345
347
The characteristics of Geo3DML are summarized as follows:
348 349
Geo3DML is based on a unified and formal representation of structural models, attribute models and
RI PT
346
hierarchical structures of geodata, including drill, cross-sections/geomaps and 3D models.
The framework of Geo3DML is based on a policy of segregation of geomodel data and visual information, and draws on the common hierarchical management of spatial data in 2D GISs. In this
351
way, users can be allowed to deal with the whole project, a geomap or some layers, and a geomodel or
352
some features.
354
The extensions of existing standards used in Geo3DML mainly entail geometry data and visual styles,
M AN U
353
SC
350
supporting boundary-based and grid-based representation and visualization for geomodels.
355 356 357
The limits of the present version of Geo3DML include:
Geo3DML provides two choices for the exchange of GeologicFeatures. Although it is flexible for data providers to choose the built-in classes or user-defined attributes, the flexibility may result in
359
difficulties of understanding the contents of geoobjects.
360
TE D
358
Not all 3D grid types are defined and supported, such as the PEBI mesh and GTP (Wu, 2004). Geo3DML now defines some widely used grid types and provides an extension mechanism for
362
generalization of the class GeoVolume (Fig. 10).
364
xml document), which is not efficient for massive amounts of data. Dividing data into blocks and
365 366
building a spatial index needs further consideration.
367 368
Only the common 3D visual information including RGBA-based colors and textures is supported.
Other visual parameters, such as the HSV color model, are not considered.
369 370
One geometry object or one array of field data is encoded in one data block (one element node in an
AC C
363
EP
361
Geo3DML now focuses on interpreted resultant geological models. Other data like geophysical data or data from other sources may not be well supported.
Evaluation criteria of geomodels are not considered. The quantitative and qualitative descriptions of the
371
quality measures of attribute models and structural models are still open. We will investigate these
372
problems further.
ACCEPTED MANUSCRIPT 373 374
The definition of geomodel-related metadata is still simple and cannot standardize the structures and contents of provenience, modeling process and products.
7. Conclusions
376
This paper introduced Geo3DML, a solution to the exchange of 3D geomodels based on the existing OGC
377
standards. The requirements for creating Geo3DML were derived from the need for interoperability of 3D
378
geodata in GSOs. Geo3DML focuses on the exchange of interpreted resultant geomodels in different
379
dimensional views, including drill, cross-sections/geomaps and 3D models. These geomodels are presented
380
based on a unified and formal method. Geo3DML considers geomodels integrally to encode all the semantic and
381
geometric information of geoobjects and their relationships, as well as visual information. Now, Geo3DML and
382
some supporting tools have been released as a data-exchange standard for 3D geomodels in the CGS.
384
SC
M AN U
383
RI PT
375
We also hope the design of Geo3DML will provide some inspiration to extend the GeoSciML available for 3D geomodels.
There have been some requirements and attentions regarding the evaluation and reuse of geomodels, such as
386
quality and complexity measures (Wellmann and Regenauer-Lieb, 2012; Wellmann et al., 2014; Pellerin et al.,
387
2015) and topologically-based geological analysis (Thiele et al., 2016). Thus, the future of Geo3DML will
388
investigate a standardized encoding method for all quantitative and qualitative geomodel-related information for
389
geomodel-based management, query and analysis.
390
EP
TE D
385
Acknowledgements
392
This work was supported by the National Natural Science Foundation of China (Grant No. 41202238 and
393
41672326) and a cooperative project of the CGS (No. 12120114056701). Herewith, we would like to thank all
394
partners and consultants for their support. We thank Oliver Raymond for his help using GeoSciML.
395
References
396
Apel, M., 2006. From 3D geomodelling systems towards 3D geosciences information systems: data model, query
397 398 399
AC C
391
functionality and data management. Comput.Geosci, 32, 222-229. Berg, R.C., Mathers, S.J., Kessler, H.H., et al., 2011. Synopsis of current three dimensional geological mapping and modeling in Geological Survey organizations. Illinois State Geol.Surv.Circular, 578.
ACCEPTED MANUSCRIPT 400 401
Boisvert, E., Brodaric, B., 2007. GroundWater Markup Language (GWML): Extending GeoSciML for Groundwater. AGU Fall Meeting. AGU Fall Meeting Abstracts.
402
Bradley, P.E., Paul, N., 2014. Comparing G-maps with other topological data structures. Geoinformatica.18:595–620.
403
Burns, K.L., 1988. Lithologic topology and structural vector fields applied to sub- surface predicting in geology. In: Proc.. of
405 406
GIS/LIS.
RI PT
404
Caers, J., Beirlant, J., Maes, M.A., 1999. Statistics for modeling heavy tailed distributions in geology: part I. methodology. Math.Geosci, 31, 391-410.
Caumon, G., Lepage, G., Sword, C.H., et al., 2004. Building and editing sealed geological model. Math.Geol, 36, 405-424.
408
Caumon, G., Collon-Drouaillet, P., Veslud, C.L., et al., 2009. Teacher’s aide: surface-based 3D modeling of geological
409
SC
407
structures. Math.Geosci, 41, 927-945.
Caumon, G., 2010. Towards stochastic time-varying geological modeling. Math.Geosci, 42, 555-569.
411
CityGML, 2012. O GC City Geography Markup Language (CityGML) Encoding Standard
413 414 415 416 417 418
, OGC document: 12-019.
Available at < http://www.opengeospatial.org/standards/citygml> (accessed 27.2.16). Collon, P., Steckiewicz-Laurent, W., Pellerin, J., et al., 2015. 3D geomodelling combining implicit surfaces and Voronoi-based remeshing: A case study in the Lorraine Coal Basin (France). Comput.Geosci, 77, 29–43. CSIRO,
2012.
Exploration:
a
glass
earth.
Available
at
<
TE D
412
M AN U
410
http://www.csiro.au/Portals/Publications/Magazines/resourceful/Issue-3/3-A-glass-earth.aspx >(accessed 27.2.16). De Kemp, E.A., 1999. Visualization of complex geological structures using 3D bezier construction tools. Comput.Geosci, 25(5), 581-597.
DGSM, 2005. Digital Geoscience Spatial Model Project Final Report (Eds. Smith, I.F.). British Geological Survey
420
Occasional Publication No. 9, 2005. Available at< http://www.bgs.ac.uk/downloads/start.cfm?id=535>(accessed
421
27.2.16).
423 424 425
AC C
422
EP
419
FE, 2005.
OpenGIS Filter Encoding Implementation Specification, Version 1.1.0, OGC document: OGC 04-095. Available
at < http://www.opengeospatial.org/standards/filter> (accessed 27.2.16). GeoSciML, 2016.
OGC Geoscience Markup Language 4.1 (GeoSciML), OGC document: 16-008. . Available at <
http://www.opengeospatial.org/standards/geosciml>(accessed 27.2.16).
426
GMD, 2007.
427
http://www.iso.org/iso/catalogue_detail.htm?csnumber=32557> (accessed 27.2.16).
428
GML, 2007. Geography markup language (GML) encoding standard, version 3.2.1, OGC document: 07-036. Available at
429
Geographic information -- Metadata -- XML schema implementation, ISO/TS 19139:2007. Available at <
(accessed 27.2.16).
ACCEPTED MANUSCRIPT
434 435 436 437 438 439 440 441 442 443 444 445
survey organisation. Comput.Geosci, 35(4), 820–835. INSPIRE, 2008. INSPIRE-Infrastructure for Spatial Information in the European Community. Available at
RI PT
433
Howard, A.S., Hatton, B., Reitsma, F., et al., 2009. Developing a geoscience knowledge framework for a national geological
Kessler, H., Mathers, S., Sobisch, H., 2009. The capture and dissemination of integrated 3D geospatial knowledge at the British geological survey using GSI3D software and methodology. Comput.Geosci 35 (6), 1311–1321.
Lemon, A. M., Jones, N. L., 2003. Building solid models from boreholes and user-defined cross-sections. Comput.Geosci, 29,
SC
432
Infrastructure. AGU Fall Meeting. AGU Fall Meeting Abstracts.
547-555.
Leopold, J.L., Sabharwal, C.L., Ward, K.J., 2015. Spatial relations between 3D objects: The association between natural
M AN U
431
Gupta, A., Schachne, A., Condit, C., et al., 2015. GeoSciGraph: An Ontological Framework for EarthCube Semantic
language, topology, and metrics. J.Vis. Lang. Comp, 27(1), 29–37.
Le, H.H., Gabriel, P., Gietzel, J., et al., 2013. An object-relational spatio-temporal geoscience data model. Comput.Geosci, 57, 104-115.
Lienhardt, P., 1994. N-dimensional generalized combinatorial maps and cellular quasi manifolds. Int.J.Comp. Geom.Appl,4, 275-324.
TE D
430
Mallet, J.L., 1997. Discrete modeling for natural objects. Math.Geol, 29 (2), 199–219.
447
Mallet, J.L., 2002. Geomodelling. Oxford University Press, New York, pp.6-24.
448
Ming, J., Pan, M., Qu, H., et al., 2010. GSIS: A 3D geological multi-body modeling system from netty cross-sections with
450 451 452
topology. Comput.Geosci, 36(6), 756–767. OneGeology,
2008.
OneGeology
-
Making
Geological
Map
Data
for
Earth
Accessible.
Available
at
AC C
449
EP
446
(accessed 27.2.16). PDF3D, 2007. Adobe® Acrobat SDK Implementation of the PDF Specification (Acrobat SDK Version 8.1). Available at
453
454
> (accessed 27.2.16).
455 456 457 458 459
Pellerin, J., Caumon, G., Julio, C., Mejia-herrera, P., et al., 2015. Elements for measuring the complexity of 3D structural models : Connectivity and geometry. Comput.Geosci, 76, 130–140. RESQML,
2012.
RESQML
Version1.1
Specifications.
Available
at
<
http://www.energistics.org/reservoir/resqml-standards/current-standards > (accessed 27.2.16). SE, 2006.
OpenGIS Symbology Encoding Implementation Specification, Version 1.1.0, OGC document: OGC 05-077r4.
ACCEPTED MANUSCRIPT
466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489
document: 05-078r4. Available at < http://www.opengeospatial.org/standards/sld> (accessed 27.2.16).
RI PT
465
OGC
Sprague, K.B., and De Kemp, E.A., 2005. Interpretive tools for 3-d structural geological modelling part ii: surface design from sparse spatial data. GeoInformatica, 9(1), 5-32. SWE, 2009.
OpenGIS SWE Common Data Model Encoding Standard, Version 2.0.0, OGC document: OGC 08-094r1.
Available at < http://www.opengeospatial.org/standards/swecommon> (accessed 27.2.16).
SC
464
SLD, 2005. OpenGIS Styled Layer Descriptor Profile of the Web Map Service Implementation Specification.
Thiele, S.T., Jessell, M.W., Lindsay, M., et al., 2016. The topology of geology 1 : Topological analysis. J.Struct.Geol, 91, 27–38.
M AN U
463
1095–1103.
USGS, 2007. Facing tomorrow’s challenges—U.S. geological survey science in the decade 2007–2017. USGS Cicular, 1309, 1-21
Wang, Z., Qu, H., Wu, Z., Yang, H., et al., 2016. Formal representation of 3D structural geological models. Comput.Geosci. 90, 10-23. WebGL
,
2014.
WebGL
Specification
Version
1.0.3,
27
October
2014.
Available
at
<
TE D
462
Sen, M., Duffy, T.R., 2005.GeoSciML: development of a generic geosciences markup language. Comput.Geosci, 31,
https://www.khronos.org/registry/webgl/specs/1.0/> (accessed 27.2.16). Wellmann, J.F., Regenauer-Lieb, K., 2012. Uncertainties have a meaning: Information entropy as a quality measure for 3-D geological models. Tectonophysics, 526-529, 207–216.
EP
461
Available at < http://www.opengeospatial.org/standards/se> (accessed 27.2.16).
Wellmann, J.F., Lindsay, M., Poh, J., et al., 2014. Validating 3-D Structural Models with Geological Knowledge for Improved Uncertainty Evaluations. Energy Procedia, 59, 374–381.
AC C
460
WFS, 2005, Web feature service (WFS), implementation standard, version 1.1.0, OGC document: 04-094. Available at< http://www.opengeospatial.org/standards/wfs> (accessed 27.2.16). Wu, L.X., 2004. Topological relations embodied in a generalized GTP model for 3D geosciences modeling system. Comput Geosci, 30, 405-410.
WVS, 2010, Web view service (WVS), Draft for candidate Paper version 0.3.0, OGC document: 09-166r2. Available at < http://www.webviewservice.org/> (accessed 27.2.16). X3D, 2004. Information technology computer graphics and image processing-Extensible3D (X3D), ISO/IEC19775:2004. Available at < http://www.web3d.org/x3d/specifications/> (accessed 27.2.16). Xu, N., Tian, H., 2009. Wire frame: a reliable approach to build sealed engineering geological models. Comput.Geosci, 35,
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
30, 419-428.
TE D
492
Zlatanova, S., Rahmanb, A.A., Shi, W.Z., 2004. Topological models and frameworks for 3D spatial objects. Comput.Geosci,
EP
491
1582-1591.
AC C
490
ACCEPTED MANUSCRIPT • • •
AC C
EP
TE D
M AN U
SC
RI PT
•
We propose Geo3DML for the exchange of 3D geomodels based on the OGC standards. Geo3DML is based on a unified representation of geomodels in different dimensional views. Geo3DML will provide inspiration to extend the GeoSciML available for 3D geomodels. Geo3DML and supporting tools have been released by the China Geological Survey (CGS).
S0098-3004(17)30271-6
DOI:
10.1016/j.cageo.2017.09.008
Reference:
CAGEO 4022
To appear in:
Computers and Geosciences
Received Date: 8 March 2017 Revised Date:
25 August 2017
Accepted Date: 15 September 2017
Please cite this article as: Wang, Z., Qu, H., Wu, Z., Wang, X., Geo3DML: A standard-based exchange format for 3D geological models, Computers and Geosciences (2017), doi: 10.1016/ j.cageo.2017.09.008. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT 1 2 3
5 6 7 8
Geo3DML: A standard-based exchange format for 3D geological models
RI PT
4
Zhangang Wang a*, Honggang Qub, Zixing Wub, Xianghong Wangb a
College of Geosciences and Surveying Engineering, China University of Mining and Technology, Beijing, 100083 China b
Development & Research Centre, China Geological Survey, Beijing, 100037, China *
(Corresponding author: [email protected])
SC
9
Abstract: A geological model (geomodel) in three-dimensional (3D) space is a digital representation
11
of the Earth’s subsurface, recognized by geologists and stored in resultant geological data (geodata).
12
The
13
addressed by developing standard-based exchange formats for the representation of not only a single
14
geological object, but also holistic geomodels. However, current standards such as GeoSciML cannot
15
incorporate all the geomodel-related information. This paper presents Geo3DML for the exchange of
16
3D geomodels based on the existing Open Geospatial Consortium (OGC) standards. Geo3DML is
17
based on a unified and formal representation of structural models, attribute models and hierarchical
18
structures of interpreted resultant geodata in different dimensional views, including drills,
19
cross-sections/geomaps and 3D models, which is compatible with the conceptual model of GeoSciML.
20
Geo3DML aims to encode all geomodel-related information integrally in one framework, including
21
the semantic and geometric information of geoobjects and their relationships, as well as visual
22
information. At present, Geo3DML and some supporting tools have been released as a data-exchange
23
standard by the China Geological Survey (CGS).
EP
TE D
increasing demand for data management and interoperable applications of geomodelscan be
AC C
24 25 26
M AN U
10
Keywords: Geo3DML; geological model; standard; data exchange
27 28
1. Introduction
29
3D geological modeling methods have matured in recent years and have been widely applied (Mallet, 1997; De
30
Kemp, 1999; Lemon and Jones, 2003; Sprague and De Kemp, 2005; Apel, 2006; Caumon et al., 2009; Caumon,
ACCEPTED MANUSCRIPT 2010; Collon et al., 2015). Geological survey organizations (GSOs) from different countries have put forward
32
3D geological modeling and mapping programs, which aim toward building a 3D geological framework that
33
provides a basic platform for full 3D cognition of the subsurface (DGSM, 2005; USGS, 2007; Kessler et al.,
34
2009; Berg et al., 2011; CSIRO, 2012; Gupta et al., 2015). In general, from the view of geological recognition
35
or data observation and collection, so-called geological data (geodata) mainly include boreholes, cross-sections,
36
geologic maps and 3D models and these data are actually specific forms of geological models (geomodels),
37
which provide an interpretative cognition and digital representation of Earth’s subsurface (Mallet, 2002). The
38
continuing progress in the generation and analytical methods of geomodels or geodata creates strong
39
expectations on agencies, especially geological survey organizations (GSOs), for exchanging and sharing
40
re-usable geodata, just as OneGeology (2008) shares 2D geoscience data via the internet.
SC
RI PT
31
It is necessary to provide standardized information encoding rules, making data providers and users process
42
the contents of geodata in a consistent way (Howard, 2009). The contents of a single geological object
43
(geoobject) generally include two important and distinct aspects: (1) human words or vocabulary-based
44
conceptual attributes; and (2) location or shape related geometric information. Currently, there are several
45
standards developed to describe the geoobject-related contents. GeoSciML (Sen and Duffy, 2005) defined very
46
basic geological data structures and document formats based on GML (2007), and GeoSciML 4.1 has now been
47
adopted as an OGC standard (GeoSciML, 2016). The geometry data type from GML can be used to represent
48
the shape of a geoobject. RESQML (2012) is mainly used for data exchange of 3D reservoir objects in the oil
49
and gas exploration industry. However, these still lack holistic geomodel representations. The purpose of
50
designing an exchange standard is more than providing a readable file format. The holistic representation of
51
geomodels encompasses the rules of space partitioning, the semantic structures of geoobjects and the
52
identification of relations between geoobjects (Wang et al., 2016), as well as other information for exchange,
53
including metadata, quality evaluation, visual information, etc. All such information can be considered as
54
geomodel-related data for reuse, and should be organized and encoded in a unified framework. Thus, Geo3DML
55
is proposed to provide a standardized representation and exchange format for 3D geomodels based on the
56
existing standards.
AC C
EP
TE D
M AN U
41
57
The remainder of this paper is structured as follows. The design principles of Geo3DML are given in Section
58
2. Section 3 introduces a unified and formal method of representing interpreted resultant geomodels in different
59
dimensional views. Section 4 introduces the framework and main components of Geo3DML. Section 5 gives
ACCEPTED MANUSCRIPT 60
information relating to supporting tools. Section 6 discusses the features and limits of Geo3DML. Section 7
61
presents our conclusions and plans for future work.
62
2. Design principles
63 Geological survey organizations
Users
Institutes & Enterprises
Software1 WVS
3D Geological Spatial Database
PDF3D
Data sharing
Government Departments
……
Other Consumers
M AN U
……
WFS
SC
Spatial Database
Geo3DML
Integration
Software3
Geo3DML
3D Geological Software2
RI PT
3D Geological Modeling and Analysis Tools
Exchange
64 65 66 67 68 69 70 71 72 73 74 75 76 77
Management platform
……
Fig. 1. Requirements for exchanging and sharing geomodels using Geo3DML for GSOs
One of the main responsibilities of GSOs is to manage processed digital geodata and to promote geodata
79
services in different applications, as shown in Fig. 1(dash line). The geodata can not only realize the conceptual
80
description of subsurface information, but can also record visual information. Geo3DML
81
encoding of geomodels in 3D geospace, is designed to transfer geodata in different stages, encompassing: (i)
82
exchanging geomodels between different software; (ii) ensuring geomodels integrated within spatial databases;
83
and (iii) sharing geomodels in a standardized service-oriented architecture, such as WFS (2005) and WVS
84
(2010), or other applications, e.g., PDF3D (2007). Then, users can download, visualize or query the geodata.
85
Thus, the principles in creating Geo3DML include:
EP
87
Platform independent and integrated data model. As a public exchange format, Geo3DML is
independent of specific geomodeling software. Moreover, Geo3DML provides a unified way to express
88
structural models and attribute models and organize the hierarchical structure of geodata, including
89 90
through standard
AC C
86
TE D
78
drills, cross-sections/geomaps and 3D models (see Section 3).
Segregation of geomodel data and visual information. The core framework of Geo3DML in Section 4.1
91
is based on a loosely coupled segregation policy, such that Geo3DML can support modular data
92
exchange. In other words, geodata can be separated or shared as needed. For example, data users can
93
obtain a complete geomodel without visual information.
ACCEPTED MANUSCRIPT 94
Reusing and extending existing standards. Geo3DML mainly uses the existing OGC standards (Table
95
1), based on GML extensions, and the conceptual model of geoobjects is derived from OGC
96
GeoSciML. Thus, Geo3DML is allowed compatibility with the existing OGC web-service
97
architectures.
Standard name Filter Encoding
Version 1.1.0
GeoSciML
Geoscience Markup Language
4.1
GMD
Geographic Information-Metadata
GML
Geography Markup Language
3.2.1
SE
Symbology Encoding
1.1.0
SWE
Sensor Web Enablement Common Data Model
2.0
100 101
Contribution in Geo3DML Extension for encoding 3D visual parameters Definition of geoobjects and their relations Extension for encoding project and geomodel metadata Extension for encoding geometry-related shape objects and field data Extension for encoding 3D visual parameters Definition of the attribute structure of geoobjects
SC
Abbreviation FE
RI PT
Table 1. Relevant standards used in Geo3DML
M AN U
98 99
Modularity and extendibility. Geo3DML uses the XML Schema language and UML to define the
102
framework (see Section 4.1), which mainly includes the seven modules shown in Fig. 2. The function
103
of each module is given in Table 2.
<
Geo3DProject
TE D
GeoMetadata
<
<
<
104 105 106 107
AC C
EP
GeoGeometry
GeoFeature
Geo3DStyle
<
GeoBasicType
<
GeoProperty
Fig.2. Modular components of Geo3DML
Table 2. Description of the Geo3DML modules No 1
Module name Geo3DProject
2
GeoMetadata
3
Geo3DStyle
4
GeoFeature
5
GeoGeometry
Description Defines the project information, and the structure of geomodels. This is the main component of the data transfer format defined by Geo3DML Defines the structure of the GeoModel metadata Defines the 3D visualization parameters and the structure of the parameter library Defines the geological features and the structure of the classes of geological objects Defines the extended structure of geometric objects stated in the GML specifications
ACCEPTED MANUSCRIPT
108 109
6
GeoProperty
Defines the extended structure stated in gmlcov:Abstract-Coverage and is used to describe the 3D property fields
7
GeoBasicType
Defines the structure of the basic data used by Geo3DML
Ease of use. A series of support tools have been developed in order to help users to check the veracity of Geo3DML files, visualize geomodels and program Read-Write interfaces.
3. Methods of representing geomodels
111
3.1 Three levels for representing geomodels
112
The formal representation of geomodels is the foundation of Geo3DML, providing a unified framework to
113
integrate geological information. Several studies attempted to represent geomodels, but their focus was on
114
constructing and editing a geological model, such as the G-maps (Lienhardt, 1994; Mallet, 2002), the Sealed
115
Geological Model (Caumon et al., 2004) and the Wire frame (Xu and Tian, 2009). The geomodels represented
116
by G-maps have complete mathematical theories to describe spatial data in any spatial dimensions, which have
117
been extended into the spatio-temporal domain (Le et al., 2013). More relevant discussions about these models
118
were given in Wu and Caumon's studies (Wu, 2004; Caumon, 2010). Mallet (2002) noted that a geomodel
119
provides an abstract digital representation of the Earth’s subsurface, composed of both a geological structural
120
model, which stores the shapes and relations of geoobjects utilizing surface-based boundary representations, and
121
an associated attribute model, used to describe the material properties of these geoobjects. Such representation
122
involves spatial partitioning under specific conditions by constructing a mapping relationship between the
123
segmented spaces and geoobjects (Wang et al., 2016). Geomodels consist of a set of exclusive and
124
interconnected geoobjects, which define their geological semantics and geometric properties, and specify
125
topological relationships and self-consistencies. Moreover, a geomodel is recognized and processed by
126
geologists, and appears in interpreted resultant geodata, such as boreholes, sections/geomaps, and 3D models
127
(Fig. 3), which are conventionally portrayed on maps or rendered in a 3D scene. These data can be used to
128
understand geological phenomenon along a line path, a surface, or in 3D space. Thus, geomodels provide
129
geological cognition of the subsurface in various dimensional views and have three levels (Fig. 4): (i) spatial
130
modeling is for modeling the geometric shapes of geoobjects, and uses vector or raster data types to portray
131
locations and graphic characteristics; (ii) object modeling is for modeling single, spatially referenced geoobjects,
132
where an object models the information about itself, particularly a representation of the geological setting and
133
concepts, essential attributes and linked geometries as spatial properties; and (iii) integrated modeling is for
134
modeling spatially related collections and repositories of geoobjects and for forming geomodels with explicit
AC C
EP
TE D
M AN U
SC
RI PT
110
ACCEPTED MANUSCRIPT spatial relations between objects and model-related information.
136 137 138 139 140 141 142 143 144 145 146 147 148 149
3D
Vertex Interface
Edge Domain
SC
Contact
East
Dimension of Geometric elements
of GeoModel
3
3 (3D model)
Body
2 (map, section) 1 (drill, traverse) The ‘
2
1
0
Interface
Edge
Vertex
Domain
Boundary
Junction
Interval
Contact
’ arrow means the ‘equivalent’ representation of elements
Object modeling
Spatially related collections (geomodel)
EP
Integrated modeling
TE D
Fig. 3. Geometric elements of geomodels in one, two and three dimensional views (Burns, 1988 and Thiele et al., 2016)
Spatially-referenced geological objects Id,name properties
AC C
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169
RI PT
Body
Dimension
150 151 152 153
1D
2D
M AN U
135
Spatial modeling
Stratum (Jurassic)
Geometric objects
Model types for interpreted resultant geological data: borehole, cross-section, geomap, 3D model ... Purpose: organizing and integrating the geomodel-related information in a unified way.
Object types for geological concepts: unit, fault, contact …
geologic
Purpose: defining data structures and vocabularies using GeoSciML
Geometry types for Boundaryrepresentation and Gridrepresentation: point, line, triangle (polygon), polyhedron ... Purpose: partitioning geopace and modeling geometric shapes of geoobjects by extending GML
Fig. 4. Three levels of geomodels in Geo3DML
170
A sample of a 3D model, composed of both a structural model and associated attribute model, is shown in Fig.
171
5. The modeling process involves the partitioning and discretization of the 3D subsurface. The boundary
ACCEPTED MANUSCRIPT surfaces of the structural model are generated based on a number of geological boundary conditions (e.g., faults
173
or horizons), and then each complete geoobject is closely formed by the corresponding surfaces. In current
174
practice, the representation of geoobjects can differ in their semantic and spatial relationships. For example (Fig.
175
5c, d and e), three possible common types of topological representations of geo-unit A can be categorized, where
176
the third method represents explicit topological relationships between A and geological boundaries. The attribute
177
model of this sample is constructed using a grid/voxel-based model, supporting discretization of the space of a
178
sealed geo-unit A (Fig. 5f and g) to generate the distribution of geological attribute fields (Caers et al., 1999;
179
Mallet, 2002; Caumon, 2010).
RI PT
172
M AN U
SC
3D Model
Stratum A
(a)
180 181
(b)
182 (c)
(d)
(e)
Grid/Voxel-based representation
AC C
183 184
EP
TE D
Surface-based representation
185 186 187 188 189 190
(f)
(g)
Fig. 5. Methods of representation of a 3D geomodel: (a) A 3D structural geomodel; (b) Wireframe representation of stratum A using triangulated irregular networks (TINs); (c) A is formed using one fully-closed surface; (d) A is composed of two components split by a fault; (e) A is topologically formed by combining boundary surfaces; (f) A discrete attribute model
ACCEPTED MANUSCRIPT embedded in stratum A using tetrahedral meshes; and (g) An attribute model using cornerpoint grids.
193
3.2 Integrated model
194
A geomodel is first and foremost a spatial data model. A full description of a spatial model involves three
195
aspects: spatial partitioning, constructive rules, and supported objects and primitives (Zlatanova et al., 2004). As
196
introduced in our work for 3D models (Wang et al., 2016), geomodels in n-dimensional views under the same
197
time and observation scales can be formally given:
198
n-GeoModel = {Gp, O, R, A} (n = 0, 1, 2 and 3)
RI PT
191 192
(1)
Gp represents geometric discretization and partitioning of subsurface space.
200
O represents geoobjects (i.e., GeologicFeatures in GeoSciML and Geo3DML). Wang et al. (2016) gave
SC
199
the constructive rules and formal definitions of the basic types of geoobjects—GeologicBoundary,
202
GeologicUnit and GeologicStructure. GeologicBoundary is a special type of GeologicStructure. It is an
203
abstract representation of all geological boundary constraints for spatial partitioning, such as a part of a
204
fault or a horizon (stratigraphic/lithological boundary) in an actual model. For a GeologicUnit object o,
205
the geometrical dimension dim(o) = n, and for a GeologicStructure, dim(o) = n-1.
R represents explicit spatial relationships between geoobjects. The formal representation of spatial
TE D
206
M AN U
201
relations is becoming increasingly important for alphanumeric query and analysis of spatial data
208
(Bradley and Paul, 2014; Leopold et al., 2015; Thiele et al., 2016). GeoSciML has defined several
209
abstract geological relations and spatial relations. Geo3DML emphasizes the topological relations and
210
directional relations, where R = {above, below, at, boundary, composition}, as defined in Geoscience
211
Spatial Framework (GSF) (DGSM, 2005) from BGS and GSIS (Ming et al., 2010). The ‘above’ or
212
‘below’ specifies a GeologicBoundary adjacent to a unique GeologicUnit above or below it. The ‘at’
AC C
EP
207
213
denotes a GeologicBoundary interpreted as a more specific GeologicStructure. The ‘boundary’
214
specifies a relationship of the bounds of a GeologicUnit composed of GeologicBoundaries. The
215
‘composition’ specifies a relationship of a GeologiUnit composed of other GeologicUnits, or the
216 217
relationship of a GeologicStructure composed of other GeologicStructures.
A represents geological attribute fields (v = F(x)) discretized and bound to a geometry object G using a
218
grid/voxel-based representation. Any scalar, vector or tensor field can be discretely stored in a primitive
219
(minimum cell) of G through four binding types as shown in Fig. 6.
220
ACCEPTED MANUSCRIPT
221 Attributes on:
Vertex
Edge
Face
Cell
RI PT
Fig. 6. Four binding types for storing the value of attribute fields in a geometric cell
Generally, n-GeoModels can be spatially represented in a unified form (Table 3) and the geological structures and units are presented by n-1- and n-Geometries, respectively. The n-Geometries can be further discretized into n-primitives, or other compatible Cartesian grids, to form the attribute models.Table 3. Spatial representations of an n-GeoModel in various dimensions n-GeoModel
Shape of GeologicBoundary
Shape of GeologicUnit
Attribute model
SC
n
(n-1-Geometry)
(n-Geometry)
(Primitives embedded in n- Geometry) Node
0
Observed Point
/
Point
1
Borehole
Point
Line
2
Cross-section/Geomap
Line
Polygon
Grid (Triangle)
3
3D Model
Surface
Body
Polyhedron (Grid3D)
M AN U
222 223 224 225 226 227 228
Segment
4. Components of Geo3DML
230
4.1 Framework
231
The top-level components of Geo3DML are shown in Fig. 7, and the core framework is composed of GeoModel
232
and Geo3DMap. The class GeoModel is designed to represent the n-GeoModels (eq. (1)). The class Geo3DMap
233
is used to define how to visualize the geomodels. Geomodels are usually created in a geological survey
234
campaign or project, and the class Geo3DProject is used to organize all the relevant GeoModels and
235
Geo3DMaps. Geo3DProject also records the project information such as location, extent and responsible
236
parties.
AC C
EP
TE D
229
237
The configuration of Geo3DMap draws on the concept of the traditional map in 2D geographic information
238
systems (GISs) and is composed of several layers. Geo3DLayer is a layer that is used for the combination of a
239
set of geoobjects (GeoFeatureClass) and their visual information (Geo3DStyle). A GeoFeatureClass is capable
240
of being visualized by a variety of Geo3DStyles, such as color or texture filling styles.
241
Geo3DML employs the fundamental class GeologicFeature from GeoSciML to model all geoobjects in a
242
geomodel. However, the data structures (semantic attributes) of GeologicFeature defined by GeoSciML are
243
fixed. Although GeoSciML-Lite (GeoSciML, 2016) provides an extendible way to allow additional attributes, it
244
is still insufficient to exchange generic feature types. Geo3DML provides another choice for defining
ACCEPTED MANUSCRIPT 245
GeoFeatures (see Section 4.2). GeoFeatureClass is a set of GeologicFeatures with the same attribute
246
configuration.
247
GeologicFeatureRelation from GeoSciML, as shown in Fig. 8. The model-related metadata, such as describing
248
the project initiative and data sources or the process of constructing the geomodels, are given by MD_Metadata
249
from GMD (2007).
five
spatial
relationships
mentioned
in
Section
250 1
Geo3DProject (f rom Geo3DML)
GeoModel
M AN U
(f rom Geo3DML)
+FeatureClasses
GeoFeatureClass (f rom Geo3DML)
1
+FeatureClass
1
choice
+Features 0..n GeoFeature
(f rom GeoSciML)
(f rom Geo3DML)
TE D
+Features 0..n GeologicFeature
+Layers
EP
1
0..n
Geo3DLayer (f rom Geo3DML)
1 1
+Styles
0..n
Geo3DStyle (f rom Geo3DML)
Fig. 7. UML class diagram: framework of top-level components
AC C
251 252 253
1
0..n
Geo3DMap
(f rom Geo3DML)
0..n
derived
+Maps
+Models
1
are
SC
0..n
1
3.2
from the
RI PT
The
ACCEPTED MANUSCRIPT GeoModelMetadata
MD_Metadata
(from Geo3DML)
(from GM D)
GeologicFeatureRelation
GeoModel
(from GeoSciML)
(f rom Geo3DML)
0..n
0..n
Geo3DProject
1
1
Geo3DMap
0..n
(f rom Geo3DML)
(f rom Geo3DML)
1 1
RI PT
1
+Layers
+FeatureClasses
0..n
1
GeoFeatureClass
1
0..n
Geo3DLayer
(f rom Geo3DML)
SC
(f rom Geo3DML)
BoundaryRelationship
DefiningStructure
BelowRelationship
AboveRelationship
CompositionRelationship
(from GeoSciML)
(from GeoSciML)
(from Geo3DML)
(from Geo3DML)
(from Geo3DML)
254 255 256
M AN U
Fig. 8. UML class diagram: metadata and relationships
4.2 Semantic and geometric attributes of geologic features
258
The procedure for modeling a single generic geoobject (Fig. 4) involves the definitions of semantic attributes,
259
geometry data and attribute fields. Fig. 9 shows the data model of geologic features.
Feature choice: GeoFeature or subclass of GeologicFeatrue from GeoSciML
0..n
TE D
257
AbstractFeature
AbstractGeometry
AbstractDiscreteCoverage
(from GM L)
(from GML)
(from GMLCov)
GeologicFeature (f rom GeoSciML)
1
0..n
1 +Shape 1
MappedFeature (from GeoSciML)
EP
+occurrence
+Features 1 GeoFeatureClass (from Geo3DML)
GeoDiscreteCoverage
AC C
GeoFeature
Schema Features FeatureType SpatialType Dimension
1
0..n
+Features
(from Geo3DML)
(from Geo3DML)
1
Fields Geometry
0..1
Geometry (from Geo3DML)
1
0..n
+ShapeProperty
PropertyFieldType : CodeList SamplingFrame : AbstractGeometry SamplingTarget : CodeList
1
1
0..1 +Schema DataRecord (from SWE)
field
260 261 262 263 264
0..n
1 1..n +field
+Fields
AbstractDataComponent (from SWE)
Coldelist of SamplingT arget VERTEX EDGE FACE CELL Coldelist of PropertyFieldT ype SCALAR VECT OR TENSOR MAT RIX
Fig. 9. UML class diagram: representation of geologic features
ACCEPTED MANUSCRIPT AbstractGeometry (from GML)
AbstractGeometricAggregate
AbstractGeometricPrimitive
Grid
(from GML)
(from GM L)
(from GML)
1
GeoGrid
1
(from Geo3DML)
+grid
1 1 +transformationMatrix AbstractCurve
AbstractSurface
DirectPosition
AbstractSolid
TransformationMatrix4x4
(from GML)
(from GM L)
(from GM L)
(from GML)
(from Geo3DML)
GeoTin
(from GM L)
(from Geo3DML)
Vertices Triangles
1
0..n
+Vertices
Vertex (from Geo3DML)
IndexNo
0..n
1
GeoVolume Vertices
+Vertices
1 GeoTetrahedronVolume (from Geo3DML)
Tetrahedrons
GeoCuboidVolume (from Geo3DML)
M AN U
Cuboids
1
0..n
+Triangles
Triangle IndexNo VertexList NeighborList
265
1
GeoCornerPointGrid (from Geo3DML)
Dimension Pillars Cells
0..n
+Cuboids
(from Geo3DML)
Cuboid
IndexNo VertexList NeighborList
(from Geo3DML)
IndexNo VertexList
TE D
Fig. 10. UML class diagram: geometry
267 268
(from GM L)
+Tetrahedrons
Tetrahedron
(from Geo3DML)
266
0..n
Solid
(from Geo3DML)
SC
LineString
RI PT
Point (from GM L)
•
There are two choices for extending attributes of a GeologicFeature in Geo3DML. One is to use the specific GeoSciML subclass of GeologicFeature. The attribute FeatureType of GeoFeatureClass
270
indicates the specific class name, e.g., GeologicUnit or Contact. Another makes use of the GeoFeature
271
from Geo3DML to add user-defined attributes. The attribute Schema of GeoFeatureClass is an instance
272
of swe:DataRecord from OGC SWE specifications (SWE, 2009), and is used for defining the structures
273
of the attribute table. The attribute Fields of GeoFeature record the value of each corresponding field
274 275
AC C
EP
269
in Schema.
•
The geometry-related shape objects and field data extend from the GML geometry and coverage
276
definitions (GML, 2007). The shape of a GeologicFeature object, i.e., MappedFeature, is associated
277
with an instance of gml:AbstractGeometry. GeoSciML only focuses on 2D geometry portrayed in
278
geologic maps. However, MappedFeature cannot support coverages and hardly covers well-known
279
geometric types used for surface-based and grid-based representations in 3D geomodeling software.
280
Geo3DML defines the Geometry class to extend MappedFeature and provides more geometric types to
ACCEPTED MANUSCRIPT 281
represent
282
tetrahedron-based model, cuboid-based model and cornerpoint grid, as shown in Fig. 10. Each
283
geometry object can be stored in binary form in XML-based documents using OGC Well-Known
284
Binary (WKB) encoding.
including
structured
and
unstructured
meshes,
such
as
the
Field data are dependent on a specific geometry object. Geo3DML defines GeoDiscreteCoverage,
RI PT
•
285
GeologicFeatures,
which inherits from gmlcov:AbstractDiscreteCoverage, to support the geological attribute model. It has
287
three main extension attributes: (1) SamplingFrame points to an associated geometry object; (2)
288
SamplingTarget refers to one of four binding types depicted in Fig. 6; and (3) PropertyFieldType refers
289
to the field type, such as scalar, vector or tensor.
SC
286
290 4.3 Geo3DStyle
292
The visual information of a geomodel is recorded in Geo3DStyles. The core model is illustrated in Fig. 11. By
293
referring to SLD (2005), CityGML (2012) and X3D (2004), the extension for encoding 3D visual parameters
294
comes from the SE (2006) and FilterEncoding (FE, 2005) specifications. FeatureTypeStyle is used to define the
295
visual parameters of geometric data, and CoverageStyle is used to define the visual parameters of the attribute
296
fields. Rule classifies the specific visual parameters (Symbolizer) based on different geological attributes using
297
Filter and ElseFilter. To support 3D visualization, Geo3DML gives common encoding of 3D visual information
298
for
299
(geo3dml:GeoSurfaceSymbolizer), and field data (geo3dml:GeoDiscreteCoverage).
(geo3dml:GeoPointSymbolizer),
AC C
EP
point
TE D
M AN U
291
line
(geo3dml:GeoLineSymbolizer),
surface
ACCEPTED MANUSCRIPT AbstractGML (from GML)
0..1
1
Geo3DStyle
(from SE)
1
CoverageStyle
0..1
(from Geo3DML)
1
(from SE)
1 0..n
0..n Rule (from SE)
Filter (from Fi lterEncodi ng)
0..1
1
1 1
ElseFilter (from SE)
0..1
1..n Symbolizer
LineSymbolizer
(from SE)
(from SE)
GeoPointSymbolizer
GeoLineSymbolizer
(from Geo3DML)
(from Geo3DML)
1
1
GeoDiscreteCoverageSymbolizer
(from Geo3DML)
(from Geo3DML)
1
1
0..2
0..2
Material
AbstractTexture
GeoSceneStyle
(from Geo3DML)
(from Geo3DML)
EP AC C
301
GeoSurfaceSymbolizer
(from Geo3DML)
AmbientIntensity DiffuseColor EmissiveColor Shininess SpecularColor Transparency
300
(from SE)
TE D
1
1
RasterSymbolizer
M AN U
PointSymbolizer
SC
(from SE)
RI PT
FeatureTypeStyle
GeoReferencedTexture
ParameterizedTexture
(from Geo3DML)
(from Geo3DML)
Fig. 11. UML class diagram: Geo3DStyle
302
5. Supporting tools
303
Geo3DML (Version 1.0) was accepted and released in December 2015 by CGS as a standardized data-exchange
304
format for 3D geomodels. It is intended that Geo3DML will ultimately encompass a broad range of
305
requirements for geomodel data interoperability. In the process of compiling the Geo3DML, five geomodeling
306
software companies in China were invited to participate in the validation from multiple geological domains,
307
including fundamental geology, hydrogeology, urban geology, mineral geology and so on. In order to facilitate
308
the easy development of Geo3DML by different users, a series of supporting tools have also been provided,
ACCEPTED MANUSCRIPT including: (1) Geo3DML Viewer is a tool to check the veracity of Geo3DML files and to visualize the data
310
running on Windows; (2) Geo3DML SDK is an open-source C++/C# program for developing Read-Write
311
interfaces for Geo3DML-based documents; and (3) other tools for data exchange between Geo3DML and other
312
well-known file formats, such as the PDF3D convertor and GoCAD convertor. As shown in Fig. 12, for
313
example, the 3D bedrock geomodel of Beijing generated by GSIS (Ming et al., 2010) was encoded using
314
Geo3DML and rendered by Geo3DML Viewer and GoCAD (Version 2011, now named SKUA), respectively.
315
The differences in visual effect between the two screenshots are due to the different settings of global scene
316
rendering parameters, such as the number of light sources and ambient light. These parameters
317
exchanged in Geo3DML.
SC
RI PT
309
are not
In addition, a protosystem was developed to visualize Geo3DML-based geomodels in a web-service
319
application implemented using WebGL (2014). All such information can be obtained from the Geo3DML
320
website www.geo3dml.cn/en.
321 322
(a)
AC C
EP
TE D
M AN U
318
RI PT
ACCEPTED MANUSCRIPT
SC
323 (b)
325
Fig. 12. The 3D bedrock geomodel of Beijing generated by GSIS (Ming et al., 2010), and rendered in: (a) Geo3DML
326
Viewer; and (b) GoCAD.
327
6. Discussion
328
At present, Geo3DML focuses on the exchange of geomodels—3D interpreted resultant geodata, serving GSOs.
329
The purpose is to provide a standardized representation and encoding method based on the existing standards,
330
especially representing geomodels associated with the conceptual model of GeoSciML. Geo3DML considers
331
geomodels integrally to organize all the information in a unified way, as well as visual information. This feature
332
is the main difference with other standards.
TE D
M AN U
324
GeoSciML is most suited for geoobject modeling (Fig. 4), supporting descriptions of geological setting and
334
concepts. However, the geometry of a GeologicFeature cannot be well modeled by the popular methods of
335
boundary-based and grid-based representations. The current extensions of GeoSciML are always implemented
336
to describe the conceptual attributes of geoobjects in specific geological domains, such as GWL (Boisvert and
337
Brodaric, 2007) for hydrology, INSIPRE (2008) for multiple disciplines, GeoSciGraph (Gupta et al., 2015) and
338
3D-GEM (Tegtmeier et al., 2014) for geotechnics. The above researches hardly mention 3D geometry related
339
extensions.
AC C
EP
333
340
RESQML is used for the data exchange of 3D reservoir objects (RESQML, 2012). The main feature of
341
RESQML is the relatively complete definitions of various types of 3D grids for numerical reservoir simulation.
342
RESQML only defines some common geoobjects, like horizon and fault, which are derived from geophysical
ACCEPTED MANUSCRIPT 343
data interpretation. So the data structures of geoobjects are different from that of GeoSciML. Meanwhile, their
344
definitions do not use or extend OGC GML or GeoSciML.
345
347
The characteristics of Geo3DML are summarized as follows:
348 349
Geo3DML is based on a unified and formal representation of structural models, attribute models and
RI PT
346
hierarchical structures of geodata, including drill, cross-sections/geomaps and 3D models.
The framework of Geo3DML is based on a policy of segregation of geomodel data and visual information, and draws on the common hierarchical management of spatial data in 2D GISs. In this
351
way, users can be allowed to deal with the whole project, a geomap or some layers, and a geomodel or
352
some features.
354
The extensions of existing standards used in Geo3DML mainly entail geometry data and visual styles,
M AN U
353
SC
350
supporting boundary-based and grid-based representation and visualization for geomodels.
355 356 357
The limits of the present version of Geo3DML include:
Geo3DML provides two choices for the exchange of GeologicFeatures. Although it is flexible for data providers to choose the built-in classes or user-defined attributes, the flexibility may result in
359
difficulties of understanding the contents of geoobjects.
360
TE D
358
Not all 3D grid types are defined and supported, such as the PEBI mesh and GTP (Wu, 2004). Geo3DML now defines some widely used grid types and provides an extension mechanism for
362
generalization of the class GeoVolume (Fig. 10).
364
xml document), which is not efficient for massive amounts of data. Dividing data into blocks and
365 366
building a spatial index needs further consideration.
367 368
Only the common 3D visual information including RGBA-based colors and textures is supported.
Other visual parameters, such as the HSV color model, are not considered.
369 370
One geometry object or one array of field data is encoded in one data block (one element node in an
AC C
363
EP
361
Geo3DML now focuses on interpreted resultant geological models. Other data like geophysical data or data from other sources may not be well supported.
Evaluation criteria of geomodels are not considered. The quantitative and qualitative descriptions of the
371
quality measures of attribute models and structural models are still open. We will investigate these
372
problems further.
ACCEPTED MANUSCRIPT 373 374
The definition of geomodel-related metadata is still simple and cannot standardize the structures and contents of provenience, modeling process and products.
7. Conclusions
376
This paper introduced Geo3DML, a solution to the exchange of 3D geomodels based on the existing OGC
377
standards. The requirements for creating Geo3DML were derived from the need for interoperability of 3D
378
geodata in GSOs. Geo3DML focuses on the exchange of interpreted resultant geomodels in different
379
dimensional views, including drill, cross-sections/geomaps and 3D models. These geomodels are presented
380
based on a unified and formal method. Geo3DML considers geomodels integrally to encode all the semantic and
381
geometric information of geoobjects and their relationships, as well as visual information. Now, Geo3DML and
382
some supporting tools have been released as a data-exchange standard for 3D geomodels in the CGS.
384
SC
M AN U
383
RI PT
375
We also hope the design of Geo3DML will provide some inspiration to extend the GeoSciML available for 3D geomodels.
There have been some requirements and attentions regarding the evaluation and reuse of geomodels, such as
386
quality and complexity measures (Wellmann and Regenauer-Lieb, 2012; Wellmann et al., 2014; Pellerin et al.,
387
2015) and topologically-based geological analysis (Thiele et al., 2016). Thus, the future of Geo3DML will
388
investigate a standardized encoding method for all quantitative and qualitative geomodel-related information for
389
geomodel-based management, query and analysis.
390
EP
TE D
385
Acknowledgements
392
This work was supported by the National Natural Science Foundation of China (Grant No. 41202238 and
393
41672326) and a cooperative project of the CGS (No. 12120114056701). Herewith, we would like to thank all
394
partners and consultants for their support. We thank Oliver Raymond for his help using GeoSciML.
395
References
396
Apel, M., 2006. From 3D geomodelling systems towards 3D geosciences information systems: data model, query
397 398 399
AC C
391
functionality and data management. Comput.Geosci, 32, 222-229. Berg, R.C., Mathers, S.J., Kessler, H.H., et al., 2011. Synopsis of current three dimensional geological mapping and modeling in Geological Survey organizations. Illinois State Geol.Surv.Circular, 578.
ACCEPTED MANUSCRIPT 400 401
Boisvert, E., Brodaric, B., 2007. GroundWater Markup Language (GWML): Extending GeoSciML for Groundwater. AGU Fall Meeting. AGU Fall Meeting Abstracts.
402
Bradley, P.E., Paul, N., 2014. Comparing G-maps with other topological data structures. Geoinformatica.18:595–620.
403
Burns, K.L., 1988. Lithologic topology and structural vector fields applied to sub- surface predicting in geology. In: Proc.. of
405 406
GIS/LIS.
RI PT
404
Caers, J., Beirlant, J., Maes, M.A., 1999. Statistics for modeling heavy tailed distributions in geology: part I. methodology. Math.Geosci, 31, 391-410.
Caumon, G., Lepage, G., Sword, C.H., et al., 2004. Building and editing sealed geological model. Math.Geol, 36, 405-424.
408
Caumon, G., Collon-Drouaillet, P., Veslud, C.L., et al., 2009. Teacher’s aide: surface-based 3D modeling of geological
409
SC
407
structures. Math.Geosci, 41, 927-945.
Caumon, G., 2010. Towards stochastic time-varying geological modeling. Math.Geosci, 42, 555-569.
411
CityGML, 2012. O GC City Geography Markup Language (CityGML) Encoding Standard
413 414 415 416 417 418
, OGC document: 12-019.
Available at < http://www.opengeospatial.org/standards/citygml> (accessed 27.2.16). Collon, P., Steckiewicz-Laurent, W., Pellerin, J., et al., 2015. 3D geomodelling combining implicit surfaces and Voronoi-based remeshing: A case study in the Lorraine Coal Basin (France). Comput.Geosci, 77, 29–43. CSIRO,
2012.
Exploration:
a
glass
earth.
Available
at
<
TE D
412
M AN U
410
http://www.csiro.au/Portals/Publications/Magazines/resourceful/Issue-3/3-A-glass-earth.aspx >(accessed 27.2.16). De Kemp, E.A., 1999. Visualization of complex geological structures using 3D bezier construction tools. Comput.Geosci, 25(5), 581-597.
DGSM, 2005. Digital Geoscience Spatial Model Project Final Report (Eds. Smith, I.F.). British Geological Survey
420
Occasional Publication No. 9, 2005. Available at< http://www.bgs.ac.uk/downloads/start.cfm?id=535>(accessed
421
27.2.16).
423 424 425
AC C
422
EP
419
FE, 2005.
OpenGIS Filter Encoding Implementation Specification, Version 1.1.0, OGC document: OGC 04-095. Available
at < http://www.opengeospatial.org/standards/filter> (accessed 27.2.16). GeoSciML, 2016.
OGC Geoscience Markup Language 4.1 (GeoSciML), OGC document: 16-008. . Available at <
http://www.opengeospatial.org/standards/geosciml>(accessed 27.2.16).
426
GMD, 2007.
427
http://www.iso.org/iso/catalogue_detail.htm?csnumber=32557> (accessed 27.2.16).
428
GML, 2007. Geography markup language (GML) encoding standard, version 3.2.1, OGC document: 07-036. Available at
429
Geographic information -- Metadata -- XML schema implementation, ISO/TS 19139:2007. Available at <
ACCEPTED MANUSCRIPT
434 435 436 437 438 439 440 441 442 443 444 445
survey organisation. Comput.Geosci, 35(4), 820–835. INSPIRE, 2008. INSPIRE-Infrastructure for Spatial Information in the European Community. Available at
RI PT
433
Howard, A.S., Hatton, B., Reitsma, F., et al., 2009. Developing a geoscience knowledge framework for a national geological
Kessler, H., Mathers, S., Sobisch, H., 2009. The capture and dissemination of integrated 3D geospatial knowledge at the British geological survey using GSI3D software and methodology. Comput.Geosci 35 (6), 1311–1321.
Lemon, A. M., Jones, N. L., 2003. Building solid models from boreholes and user-defined cross-sections. Comput.Geosci, 29,
SC
432
Infrastructure. AGU Fall Meeting. AGU Fall Meeting Abstracts.
547-555.
Leopold, J.L., Sabharwal, C.L., Ward, K.J., 2015. Spatial relations between 3D objects: The association between natural
M AN U
431
Gupta, A., Schachne, A., Condit, C., et al., 2015. GeoSciGraph: An Ontological Framework for EarthCube Semantic
language, topology, and metrics. J.Vis. Lang. Comp, 27(1), 29–37.
Le, H.H., Gabriel, P., Gietzel, J., et al., 2013. An object-relational spatio-temporal geoscience data model. Comput.Geosci, 57, 104-115.
Lienhardt, P., 1994. N-dimensional generalized combinatorial maps and cellular quasi manifolds. Int.J.Comp. Geom.Appl,4, 275-324.
TE D
430
Mallet, J.L., 1997. Discrete modeling for natural objects. Math.Geol, 29 (2), 199–219.
447
Mallet, J.L., 2002. Geomodelling. Oxford University Press, New York, pp.6-24.
448
Ming, J., Pan, M., Qu, H., et al., 2010. GSIS: A 3D geological multi-body modeling system from netty cross-sections with
450 451 452
topology. Comput.Geosci, 36(6), 756–767. OneGeology,
2008.
OneGeology
-
Making
Geological
Map
Data
for
Earth
Accessible.
Available
at
AC C
449
EP
446
453
454
> (accessed 27.2.16).
455 456 457 458 459
Pellerin, J., Caumon, G., Julio, C., Mejia-herrera, P., et al., 2015. Elements for measuring the complexity of 3D structural models : Connectivity and geometry. Comput.Geosci, 76, 130–140. RESQML,
2012.
RESQML
Version1.1
Specifications.
Available
at
<
http://www.energistics.org/reservoir/resqml-standards/current-standards > (accessed 27.2.16). SE, 2006.
OpenGIS Symbology Encoding Implementation Specification, Version 1.1.0, OGC document: OGC 05-077r4.
ACCEPTED MANUSCRIPT
466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489
document: 05-078r4. Available at < http://www.opengeospatial.org/standards/sld> (accessed 27.2.16).
RI PT
465
OGC
Sprague, K.B., and De Kemp, E.A., 2005. Interpretive tools for 3-d structural geological modelling part ii: surface design from sparse spatial data. GeoInformatica, 9(1), 5-32. SWE, 2009.
OpenGIS SWE Common Data Model Encoding Standard, Version 2.0.0, OGC document: OGC 08-094r1.
Available at < http://www.opengeospatial.org/standards/swecommon> (accessed 27.2.16).
SC
464
SLD, 2005. OpenGIS Styled Layer Descriptor Profile of the Web Map Service Implementation Specification.
Thiele, S.T., Jessell, M.W., Lindsay, M., et al., 2016. The topology of geology 1 : Topological analysis. J.Struct.Geol, 91, 27–38.
M AN U
463
1095–1103.
USGS, 2007. Facing tomorrow’s challenges—U.S. geological survey science in the decade 2007–2017. USGS Cicular, 1309, 1-21
Wang, Z., Qu, H., Wu, Z., Yang, H., et al., 2016. Formal representation of 3D structural geological models. Comput.Geosci. 90, 10-23. WebGL
,
2014.
WebGL
Specification
Version
1.0.3,
27
October
2014.
Available
at
<
TE D
462
Sen, M., Duffy, T.R., 2005.GeoSciML: development of a generic geosciences markup language. Comput.Geosci, 31,
https://www.khronos.org/registry/webgl/specs/1.0/> (accessed 27.2.16). Wellmann, J.F., Regenauer-Lieb, K., 2012. Uncertainties have a meaning: Information entropy as a quality measure for 3-D geological models. Tectonophysics, 526-529, 207–216.
EP
461
Available at < http://www.opengeospatial.org/standards/se> (accessed 27.2.16).
Wellmann, J.F., Lindsay, M., Poh, J., et al., 2014. Validating 3-D Structural Models with Geological Knowledge for Improved Uncertainty Evaluations. Energy Procedia, 59, 374–381.
AC C
460
WFS, 2005, Web feature service (WFS), implementation standard, version 1.1.0, OGC document: 04-094. Available at< http://www.opengeospatial.org/standards/wfs> (accessed 27.2.16). Wu, L.X., 2004. Topological relations embodied in a generalized GTP model for 3D geosciences modeling system. Comput Geosci, 30, 405-410.
WVS, 2010, Web view service (WVS), Draft for candidate Paper version 0.3.0, OGC document: 09-166r2. Available at < http://www.webviewservice.org/> (accessed 27.2.16). X3D, 2004. Information technology computer graphics and image processing-Extensible3D (X3D), ISO/IEC19775:2004. Available at < http://www.web3d.org/x3d/specifications/> (accessed 27.2.16). Xu, N., Tian, H., 2009. Wire frame: a reliable approach to build sealed engineering geological models. Comput.Geosci, 35,
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
30, 419-428.
TE D
492
Zlatanova, S., Rahmanb, A.A., Shi, W.Z., 2004. Topological models and frameworks for 3D spatial objects. Comput.Geosci,
EP
491
1582-1591.
AC C
490
ACCEPTED MANUSCRIPT • • •
AC C
EP
TE D
M AN U
SC
RI PT
•
We propose Geo3DML for the exchange of 3D geomodels based on the OGC standards. Geo3DML is based on a unified representation of geomodels in different dimensional views. Geo3DML will provide inspiration to extend the GeoSciML available for 3D geomodels. Geo3DML and supporting tools have been released by the China Geological Survey (CGS).