You are on page 1of 8

Proceedings of the Open source GIS - GRASS users conference 2002 - Trento, Italy, 11-13 September 2002

Construction and Visualization of Three Dimensional
Geologic Model Using GRASS GIS
Shinji Masumoto*, Venkatesh Raghavan**, Tatsuya Nemoto*, Kiyoji Shiono*
* Department of Geosciences, Osaka City University, 3-3-138 sugimoto, sumiyoshi-ku, Osaka 558-8585,
Japan, e-mail
** Media Center, Osaka City University, 3-3-138 sugimoto, sumiyoshi-ku, Osaka 558-8585 , Japan,

1 Introduction
Recently, the need of the geologic information has been rising in many fields such as
environmental geology, disaster mitigation, and urban geological applications. For these
fields, it is effective to provide geologic information as a three dimensional(3-D) model
that can be generated and visualized in general purpose GIS software. The present work
aims at introducing a basic theory, implementing methodology and algorithms for 3-D
modeling and visualization of geologic model using the Open Source GRASS GIS
environment. 3-D geologic model is constructed from the boundary surfaces of geologic
units and the logical model of geologic structure. The algorithms for construction and
visualization of the proposed model are based on the geologic function g. The geologic
function g assigns a unique geologic unit to every point in the objective 3-D space. The
boundary surface that divides the objective space into two subspaces, were estimated
using data from field survey. The logical model showing the hierarchical relationship
between these boundaries surfaces and geologic units can be automatically generated
based on the stratigraphic sequence and knowledge of geologic structures. Based on these
algorithms, 3-D geologic model can be constructed virtually on GRASS GIS. Applying
this model, various geologic surface and section models can be visualized in GRASS GIS
environment. Further, “Nviz” was used for dynamic visualization of geologic crosssections and generation of animated image sequences.

2 Basic theory and algorithms
2.1 Geologic function and logical model of geologic structure
Let a 3-D subspace Ω be a survey area and suppose that the area Ω is composed of n
geologic units that are disjoint:
b1 ∪ b2 ∪ ⋅⋅⋅ ∪ bn = Ω ,
bi ∩ bj = φ ( i ≠ j ) .
In order to realize a 3-D geologic visualization in the GIS environment, we have
introduced a concept of a geologic function g which assigns a unique geologic unit to
every point in the 3-D space Ω [1] [2].
g : Ω → B,

where B = {b1, b2, ..., bn} .

Fundamentals of the geologic function g is explained using a simple geologic structure
composed of three geologic units as shown in Fig. 1(a). Three geologic units b1, b2 and b3
are defined by two boundary surfaces S1, and S2 which divide Ω into two subspaces as

 +  Where hi ( d i ) =  S i . M00 = S1− ∩ S2− . bn are defined by surfaces. the logical model of geologic structure [1].. the geologic unit lies above the corresponding boundary surface. M01 ⊂ b3 . b) logical model (+1. M11 = S1+ ∩ S2+ .. Let Md1d2.Construction and Visualization of 3-D Geologic Model Using GRASS GIS 2 (a) S 2 (M11) (b) b3 (M01) b2 (M10) b1 (M00) S1 b1 b2 b3 S1 −1 +1 0 S2 −1 −1 +1 (c) minset M00 M01 M10 M11 unit b1 b3 b2 b3 Figure 1: Basic elements of a geologic model.dn-1 = h1(d1) ∩ h2(d2) ∩ . M11 ⊂ b3 .. Si. 1(b)..... .dn-1 be a minset defined by . b2 = M10 ..). −1. b3 = S2+ . ∩ hn-1(dn-1) .di. b2 = S1+ ∩ S2− . As the geologic units b1. i.di.. Md1d2. b1 = M00 . and 0 . M10 = S1+ ∩ S2− ... they can be expressed in a “minset standard form” [3]. four minsets can be defined as follows... the geologic unit lies below the corresponding boundary surface. b2. Sn−1 in the 3-D space Ω... no specific relation with the surface. M00 ⊂ b1 . ... It is evident that each minset is included in only one of geologic units as shown below..e. The above equations and table define the relation between geologic units and their boundaries.. S2. − S i . respectively. and c) relational code table. b3 = M01 ∪ M11 . a) relation between geologic units and surfaces in geologic section. M10 ⊂ b2 .. b1 = S1− ∩ S2− . M01 = S1− ∩ S2+ . d i = 1 . ∩ hi (di) ∩ . where Si+ and Si− give subspaces that lie above and below the surface Si. . d i = 0  In the case of Fig. The minset is a minimum subspace that is divided by the boundaries S1. 1(a).. These equations can be expressed in a tabular form as shown in Fig.. The minset standard forms can be derived for the geologic units as follows.

and d) geologic profile. Therefore. the point lying on the boundary and the region under the boundary have the same geologic unit name.Shinji Masumoto. z): g (x. The relation between minsets and geologic units can be expressed by a function g1 from a class of minsets I into B: g1: I → B .. a point on the boundary surface is included in the lower side of the boundary. z) = g1 ( g2 (x. y. Further.. z) in a space Ω. the boundary surface is shown as a boundary line. b) draw lines with the color of the boundary surface. In the vertical section. This function g1 can be represented by the relational code table shown in Fig. a) 3-D geologic model image in vertical section. Consequently. Tatsuya Nemoto.. The function g: Ω → B defines a rule to assign a unique geologic unit to every point in a 3-D space Ω. for a point P(x. and the minset is shown as a polygon of . and c) draw the lines with mask. respectively.dn-1 can be assigned a value of di = 1 or di = 0 depending on whether P(x. y. Venkatesh Raghavan. 2. z)) .di. 1(c). c) fill the polygon with the color of the upside line. b) geologic boundary surface judgement.. a convolution of functions g1: I → B and g2: Ω → I provides a rule to define the geologic unit that includes a given point P(x. z) falls in Si+ or Si −. y. When the point on the boundary was input to the geology function g. (a) (b) S3 (c) non-display (masked) S2 S1 display Figure 3: Flow of the geologic boundary surface generation. y. y.2 Geologic profiles and geologic boundary surfaces According to the definition of geologic function g. It was defined that the point belongs to the minset under the boundary by the function g2. This correspondence between every point in Ω and minsets is expressed by a function g2 : g2: Ω → I . a) 3-D geologic model image in vertical section. Kiyoji Shiono (a) (b) S3 (c) 3 (d) minset S2 minset S1 minset minset Figure 2: Flow of the geologic profile generation. a minset Md1d2.

3). For example. and Mo is a minset surface calculated before.relcass” command and options.0. Applying this method. the geologic profile can be drawn by assigning the polygon of the minset with the color corresponding to the geologic unit name of the upside boundary line (Fig. Sm is the raster file name of these boundary surface.0. step 3.0)∗20 (where 2k is the real value of 2k). This judgment can be drawn by comparing the geologic unit name on either side of the boundary surface using the geologic function g.mapcalc command. the raster files that are defined as the geologic category number on the boundary surface can be calculated by the geologic function g. This polygon and the upper boundary line bounding this polygon have the same geologic unit name.0.1. . various geologic surface and section models without vertical section can be visualized in GRASS GIS environment. generate the binary code to arrange the boundary surfaces in their ascending order. The following 3 steps for each grid cell of the objective surface So can express this rule. 4).0)∗2m-1 + if(So−S2. obtain the relation of the height between the surface So and every boundary surfaces (if lower and on surface then set to “0”. convert the binary code into an integer number to create the respective cell value...0. it is assumed that the Si is a elevation raster file of the boundary surface. Finally. + if(So−Sm..reclass” and the reclassification table converted from relational code table. . > r. Si.0)∗2m-2+ . > r. + if(So−Si. Consequently.. it is assumed that Lm is a file name of the reclassification table. or if higher then set to “1”).0)∗2m-i + . This geologic function g has been modified to create the geologic category raster file of the input objective raster surface (Fig. The reclassification table file can be converted into integer from binary number of the relation code table. it is assumed that m is a number of the boundary surface.. and. The geologic category number raster files Gi can be calculated by the following. the geologic features along the objective surface can be visualized with the raster files Go and So using “d. . Removing the area where upper and lower sides of the boundary surface are the same geology can show the geologic boundary surfaces that actually exist as a boundary of the geologic unit (Fig. The minset raster surface Mo of the objective surface So can be calculated by the following equation of the r.. the boundary surface that divides a space Ω into two subspaces is not a geologic boundary surface. Determination of the function g2 is implemented using a raster map calculation function “r.. In the function g. S2.1. “d.mapcalc”.1. For the geologic profile filled by the color of geologic unit..rast”.mapcalc Mo = if(So−S1.1. 2). and Li is a minset raster file for Si. a simple rule for the calculation can be constructed according to the number of boundary surfaces. 3-D geologic model can be constructed virtually by implementing the geologic function g on GRASS GIS. The function g1 is implemented to represent the geologic category map on the objective surface using the reclassification function “r. The function g2 is constructed from the raster data defining the boundary surface elevation including topographic surface. For example. Here. “<” is a redirection of unix).. can be accomplished by following “r.. step 2. step 1.4 Construction and Visualization of 3-D Geologic Model Using GRASS GIS the minset section surrounded by the multiple boundary line. In practice.. 3 Geologic modeling on GRASS GIS environment Based on the suggested algorithms..3d” and “Nvis”. and S1.reclass input = Mo output = Go < Lm (where. The geologic category map Go of objective surface So.

6. + if(Si−Sm.1.0)∗2m-2+ ...reclass input = Hi output = Ui < Lm .Shinji Masumoto. the mask files must be defined to all bounadry surfaces. 3-D geologic voxel model was constructed based on the geologic function g.reclass) Figure 4: Flow of the 3-D geologic modeling using geologic function. Northeast Japan (Fig. For drawing the geologic boundary surfaces..mapcalc command as follows. Venkatesh Raghavan. For example.1. the mask file Mai can be calculated by the r. Kiyoji Shiono Objective Surface (So) Minset Surface (Mo) Geologic Surface (Go) Surfaces Operation Rule Relational Code Table (Lm) S2 (M00) 0 So minset M 01 S1 S1 S2 Function g2 (r.1. Finally. S2. Examples of the horizontal and vertical sections are presented in Fig. For the geologic profile.mapcalc Li = if(Si−S1.reclass input = Li output = Gi < Lm . the surfaces Gi are established to the surface color and the “T” is specified to paint section by the color of the upper line in the cutting-plane of Nviz.1.1. Therefore. Examples of the voxel model visualization are presented in Fig. 9 using 3-D geologic model..0)∗2m-1 + if(Si−S2.1..0..0)∗2m-2+ .0)∗20 (where 2k is the real value of 2k). The logical model of this area is shown in Table 1. To visualize the geologic boundary surfaces. + if(Si−Si. The geologic category number raster file Ui for the upper side of the surface can be calculated by the following. The surfaces were estimated by Horizon2000[4] using data extracted from geologic map (Fig. > r.mapcalc Mai = if(Gi−Ui .1.0)∗20 (where 2k is the real value of 2k).1.. In addition.0.1.0)∗2m-i + .1) . These mask files can be generated by comparing the geology on either sides of the boundary. 4 Case study The study area is located in Honjyo region of Akita Prefecture. .1.1.0. 8. it is assumed that Hi is a minset raster file for the upper side of the surface Si.0.. The change of this equation means that the boundary surface is attributed to the upper side minset in the function g2.. the calculation of the upper side files are only necessary. + if(Si−Si. Tatsuya Nemoto. The surface geologic map with these boundary surfaces are presented in Fig. + if(Si−Sm.. The geologic boundary surfaces S1. The forementioned steps must be done to all of the boundary surfaces. > r.. > r. 5). > r. 7.mapcalc) 5 (M01) 1 (M10) 2 (M11) 3 b1 b3 b2 b3 Function g1 (r.1. 10 using Vis5D software[6].[5]).mapcalc Hi = if(Si−S1.0. > r. The lower side files were already calculated as the files Gi..0)∗2m-i + . the surfaces Si and Mai are set to the topography and mask respectively in the surface panel of Nvis.0)∗2m-1 + if(Si−S2. and S5(=DEM) are shown in Fig.1.

DEM) Figure 7: Geologic boundary surfaces. Funakawa F. (c) Surface S5(DEM)..5km. and (b) Surface S1 and S5. (8.. [5]) S4 −1 −1 −1 −1 +1 +1 S5 0 0 0 0 −1 +1 (a) (b) (c) (d) Table 1. boundary surfaces. and S5. (b) Surface S2. Gongenyama F.7X6. Onnagawa F.Construction and Visualization of 3-D Geologic Model Using GRASS GIS 6 Study Area Osaka Figure 5: Location map of the study area. S4. Alluvial Deposits α (air) S1 −1 +1 +1 +1 0 0 S2 0 −1 +1 +1 0 0 S3 0 0 −1 +1 0 0 Figure 6: Geologic map of the study area. (a) Surface S1. Logical model of the geologic structure.. (S1.. Tentokuji F. .

(a) (c) (d) (b) (e) Figure 9: Examples of the horizontal and vertical sections using 3D geologic model. Venkatesh Raghavan. (a) (b) Figure 10: Examples of the voxel model visualization using Vis5D software. Tatsuya Nemoto. 7 . bottom. middle. Kiyoji Shiono Figure 8: Geologic map with boundary surfaces. −500m). (a) Horizontal sections (top. (a) Geologic sections. and (b)~(e) Vertical sections. 0m. DEM.Shinji Masumoto. and (b) Volume model of geologic units.

To overcome these limitations the theory and functions that reflect the characteristics of geologic surfaces need to be further investigated. In this regard. Horizon2000:Revised Fortran Program for Optimal Determination of Geologic Surfaces Based on Field. A computerized geological mapping system based on logical models of geologic structures. Nonrenewable Resources 2.8 Construction and Visualization of 3-D Geologic Model Using GRASS GIS 5 Discussion In most cases. Applied algebra for the computer sciences. Scale 1:50. A. S. generation of boundary surfaces and logical model are basically based on the results of geologic field survey data. . Englewood Cliffs. 1977. [2] Shiono. S. Geology of the Honjyo District. Ikebe Y. [4] Gill. double-valued function such as over-folded structure) can not be readily generated in the present system. On formal expression of spatial distribution of strata using boundary surfaces -C1 and C2 type of contact-. 1993. [3] Shiono K. Paul B. M.ssec.J.. pages 229-249. Masumoto S. Geoinformatics. 1994. Masumoto. There still remain some limitations in this algorithm and model. K. pages 140-147. Vis5D. Kellum J. Geoinformatics. research and software development for 3-D visualization of geologic structure are oriented towards solid modeling. the proposed algorithm and the utilized system in the present paper is unique.. Therefore.: PrenticeHall. . the algorithm does not exert a heavy load on the computer system. 5. this model is flexible in comparison with the solid modeling method.html. The applied algorithm has successfully generated a 3-D geologic model utilizing the geologic function g based on the relationships between the geologic units and their boundary surfaces. The boundary surfaces of geologic model and objective surface that are difficult to represent as GIS raster layers ( 1976. pages 223-232. Masumoto. Fujioka K. 2001. 1998. M.. Wadatsumi K... References [1] Sakamoto.. Sakamoto. 12. http://www.. N. [6] Hibbard B. Noumi Y. as it uses no specialized solid modeling technique for generating geologic model.000. [5] Osawa M. Sakamoto M.. Quadrangel Series.wisc.. Takayasu T. Further. Geological Survey of Japan... Shiono.. Since the construction of geologic map is based on a virtual model. K.