3D GIS空间数据建模：国外经典教材解析

5星 · 超过95%的资源需积分: 10 152 浏览量更新于2024-07-27 收藏 5.26MB PDF 举报

"Spatial Data Modelling for 3D GIS" 是一本由 Alias Abdul-Rahman 和 Morakot Pilouk 合著的书籍，主要探讨了3D地理信息系统（3D GIS）中的空间数据建模技术。这本书是国外的最新经典教材，涵盖了与3D GIS相关的空间数据模型和理论。在3D GIS中，空间数据建模是关键组成部分，它涉及如何有效地存储、管理和分析三维空间对象。传统二维GIS中的数据模型，如矢量和栅格模型，通常不足以充分表达复杂的空间关系和三维特性。因此，3D GIS引入了新的数据结构和方法来处理立体空间信息，如多边形网络、体素模型、TIN（不规则三角网）以及点云数据模型等。 1. **3D矢量模型**：3D矢量模型通过扩展传统的点、线、面概念，引入了三维点、三维线串和三维多边形，用于表示建筑物、地形和其他三维实体。这些实体可以是封闭的几何体，也可以是开放的结构，如桥梁或隧道。 2. **体素模型**：体素类似于三维空间的像素，是3D GIS中的一种基础数据结构。它通过三维网格来表示空间，每个单元代表一个具有特定属性的小体积。体素模型适合于连续分布的三维数据，如地层结构或大气现象。 3. **TIN（不规则三角网）模型**：TIN模型基于一系列相互连接的三维三角形来构建表面，适用于不规则地形和建筑表面的表示。TIN模型允许高精度的地形起伏模拟，并能有效处理复杂的地形特征。 4. **点云数据模型**：点云由大量三维坐标点组成，每个点可能携带颜色、反射率等属性信息。这种模型常用于激光雷达扫描数据的处理，可以精确重建现实世界中的物体表面。在3D GIS中，有效的空间数据建模不仅需要理解这些基本的数据结构，还需要掌握如何进行数据集成、空间查询和分析、可视化以及数据交换标准，如CityGML和3D GML。这些标准允许不同系统之间共享3D空间信息。此外，书中的内容可能还会涉及3D GIS在城市规划、环境模拟、灾害管理、地质勘探等领域的应用，以及如何利用3D GIS技术解决实际问题。作者可能会深入讨论数据质量控制、数据优化、性能优化等话题，以确保3D GIS系统的高效运行。 "Spatial Data Modelling for 3D GIS" 是一本深入探讨3D GIS中空间数据建模理论和技术的教材，对于理解并掌握3D空间信息的管理和应用具有重要的指导价值。

6 Chapter 1

location (see Figure 1.1a). DTM cannot accommodate a 3D (solid) object,

or a surface with multiple height values at a given planimetric location (see

Figure 1.1b and Figure 1.1c, respectively).

Although raster-based systems which could be regarded as 3D GISs are

available, they may not be able to maintain the knowledge about reality

available in the original data set. This knowledge may be lost because of

problems in resolution and resampling. As a remedy, the original data set

would have to be stored separately from the model, for example, for:

• recreating the model if the result proves to be unsatisfactory because of

unsuitable mathematical definition

• creating another model with different resolution

• merging with another data set to create a new model

• archiving as a reference to, or evidence of, the model.

These activities imply the need to store original data in an appropriate

structure ready for future use. Necessary information about the data should

be attached to each data element. In DTM for instance, information that a

line is a breakline should be kept because it will have an impact on the in-

terpolation. Similarly, other information can be attached which influences

data handling strategies.

Since neither CAD nor GISs can at present fulfil the requirements of earth

science applications, further research and development of a 3D GIS would

seem appropriate.

Who needs 3D GIS?

As in the popular 2D GIS for 2D spatial data, 3D GIS is for managing 3D

spatial data. Raper and Kelk (1991), Rongxing Li (1994), Förstner (1995),

and Bonham-Carter (1996) present some of the three dimensional applica-

tion areas in GIS, including:

• ecological studies

• environmental monitoring

• geological analysis

• civil engineering

• mining exploration

• oceanography

• architecture

• automatic vehicle navigation

• archaeology

INTRODUCTION 7

Objects with known or

well-defined spatial ex-

tent, location and prop-

erties

Objects with unknown

or not well-defined spa-

tial extent, location and

properties

Fig. 1.2 Two types of real world

objects with respect to their spatial

extent.

• 3D urban mapping

• landscape planning

• defence and intelligence

• command and control

The above applications may pro-

duce much more useful information

if they were handled in a 3D spatial

system, but 3D spatial objects on

the surface and subsurface appear to

demand more complex solutions

(e.g. in terms of modelling, analysis,

and visualization) than the existing

systems can offer.

1.3 The Need for 3D Spatial

Data Modelling

In addition to the problem of creating a system capable of offering 3D

modelling and functionality, there is a further problem concerning the type

of 3D model chosen as the basis for 3D GIS. The model contains knowl-

edge about reality, so we consider below the types of real world objects it

must represent. Two kinds of real world objects may be differentiated in

terms of prior knowledge about their shapes and location, as shown in Fig-

ure 1.2. In reality, objects from the two categories coexist. Traditional GIS

models the objects of each category independently with the result that two

separate kinds of systems or subsystems have been developed.

Raper (1989) has also defined these two categories of objects. The first

category, regarded as ‘sampling limited’, is for objects having discrete

properties and readily determined boundaries, such as buildings, roads,

bridges, land parcels, fault blocks, perched aquifers. The second category,

known as ‘definition limited’, is for objects having various properties that

can be defined by means of classification, using property ranges. For ex-

ample, soil strata may be classified by grain-size distribution; moisture

content, colloid or pollutant in the water by percentage ranges; carbon

monoxide in the air by concentration ranges, and so forth. Molenaar

(1994a) regards these objects as ‘fuzzy spatial objects’.

Separate modelling of these two categories of objects tends to contradict

the reality, which leads to difficulties in representing their relationships.

Such a question as, ‘how many of the people working in a 50-storey office

8 Chapter 1

building are affected by polluted air generated by vehicles in nearby streets

during rush hours?’ cannot be answered until the two separate models are

combined, as shown in Figure

1.3. Modelling them together

with more accurate represen-

tation of their relationships in

the 3D environment requires

the integrated 3D modelling.

Note also that the properties

of an object may be well de-

fined in some specific dimen-

sions and ill defined in others.

For example, given a DTM

data set representing a sur-

face, the planimetric extent of

regions at the elevation of 100

metres above mean sea level

cannot be defined until the re-

sult of interpolation based on a mathematical definition (for example, lin-

ear interpolation) is obtained. That is to say, although the spatial extent of

this region may be known in the z-dimension, the spatial extent in plani-

metry (x, y) has still to be discovered. The model must contain the aspect

allowing the appropriate operation, such as interpolation or classification,

if the required description of the properties of an object is to be obtained.

Apart from the problem of the separate modelling of the two types of ob-

jects, there remains the further problem of the separate modelling of an ob-

ject’s components. These components are relief and planar geometry asso-

ciated with thematic properties. This separation has resulted in

independent systems and data structures, DTM and 2D GIS, respectively.

The consequences are data redundancy, which may lead to uncertainty

when the two data sets are combined and only one data set has been up-

dated.

DTM can facilitate several GIS analyses and visualization taking into ac-

counts the third dimension. The spatial information stored in DTM and in

GIS, however, can only be related through coordinates. This implies that

relationships between different components may not be properly repre-

sented because of metric computation instead of topology. To overcome

this, information derived from DTM must be converted into a form GIS

can recognize. For example, information about a slope or height zone must

first be converted into a thematic layer of GIS for further overlaying before

Fig. 1.3 An example of two types of real

world objects

INTRODUCTION 9

the spatial analysis can be carried out. Imagine having information about

the relief, planimetry and themes integrated into one model, so that con-

version of such information as slope, height zone and so forth were no

longer necessary. Such a question as, ‘which land parcels are subject to

one-metre flooding?’ could be answered from one model. Integrated mod-

elling of this kind is evidently also required for 3D GIS.

1.4 Problems Associated with Spatial Modelling

Establishing a 3D GIS while taking into account the integration of the nec-

essary components and different types of objects requires the solution of

the following problems related to the spatial model representing reality:

1) Design of a spatial model

• design of an integrated data model, or a scheme, permitting the deri-

vation of a unified data structure capable of maintaining all the com-

ponents of the geometric representation of real world objects, whether

obtained from direct measurements or from derivations, in the same

database. Each geometric component must be capable of representing

a real world object differently understood by different people.

2) Construction of a spatial model

• development of appropriate means and methods for 3D data acquisi-

tion;

• coordinate transformation into common georeferencing when differ-

ent components are to be included into one database;

• development of a data structuring method that unites the data from

various inputs of multi sources into an integrated database capable of

being maintained by a single database management system;

• design of thematic classes to organize representation of real world ob-

jects with common aspects into the same category;

• solving the uncertainty arising from discrepancies from different data

sets during the integration process and converting the uncertainty into

a ‘data quality’ statement to be conveyed to the end user.

3) Utilization of a spatial model

• utilization of existing components, such as 2D data and DTM (back-

ward compatibility) and preparation of those components for future

incorporation into the higher-dimension model (forward compatibil-

ity) to save the costs of repeating data acquisition.

for 3D GIS

10 Chapter 1

• development of additional spatial operators and spatial analysis func-

tions;

• development of maneuverable graphic visualization permitting the se-

lection of appropriate viewpoints and representation enabling conven-

ient, adequate uncovering of the details of objects stored in the data-

base;

• design of 3D cartographic presentation of information, including

name placement, symbol, generalization, etc.;

• design of a user interface and query language allowing users access to

the integrated database;

• development of a spatial indexing structure that speeds up data re-

trieval and storage processes for the integrated database, including

specific (database) views for each user group and guidelines keeping

these views updated according to the core database;

• development of tools for navigating among different models stored in

databases at different sites and computing platforms.

4) Maintenance of spatial model

• design of updating procedures, including the development of consis-

tency rules ensuring the logical consistency and integrity of the inte-

grated database, especially during the updating process.

1.5 Previous Work

The status and progress of research in the 3D GIS field within the scope of

this monograph and the identification of solutions and remaining problems

are made clear from the following review of previous work.

The development of data models for a 3D GIS has branched in two direc-

tions. The first is the full 3D approach that looks directly into the design of

a data model suitable for 3D GIS. Molenaar (1989) proposes a formal data

structure (FDS) for a 3D vector map which may be regarded as a generali-

zation of the 2D version of FDS. Shibasaki and Shaobo (1992), Rikkers

(1994), and Wang (1994) have reported

experimental use of 3D FDS.

The second approach comes from the viewpoint referred to as the ‘integra-

tion of DTM and GIS’. DTM became a discipline in its own right in the

late 1950s (Miller and Laflamme, 1958). Fritsch (1990) has recognized the

work of Makarovic (1977) as a proposer of this integration. Males (1978)

though not addressing the integration issue, demonstrated the use of a

et al. (1993), Bric (1993), Bric et al.,

剩余289页未读，继续阅读

lcybb123

粉丝: 1
资源: 9

3D GIS空间数据建模：国外经典教材解析

Spatial Data Modelling for 3D GIS.rar

3D.Global.Spatial.Data.Model.Foundation.Of.The.Spatial.Data.Infrastructure

gis.spatial.data.application.development.documents_gis

讲解nc_srs = osr.SpatialReference(nc_dataset.variables['data'].spatial_ref)

gis二次开发 平移的代码 C#语言

File "_qhull.pyx", line 1846, in scipy.spatial._qhull.Delaunay.__init__ File "_qhull.pyx", line 358, in scipy.spatial._qhull._Qhull.__init__

hull = ConvexHull(proj_points) File "qhull.pyx", line 2431, in scipy.spatial.qhull.ConvexHull.__init__ File "qhull.pyx", line 279, in scipy.spatial.qhull._Qhull.__init__ ValueError: No points given

最新资源

gis二次开发平移的代码 C#语言

File "_qhull.pyx", line 1846, in scipy.spatial._qhull.Delaunay.init File "_qhull.pyx", line 358, in scipy.spatial._qhull._Qhull.init

hull = ConvexHull(proj_points) File "qhull.pyx", line 2431, in scipy.spatial.qhull.ConvexHull.init File "qhull.pyx", line 279, in scipy.spatial.qhull._Qhull.init ValueError: No points given