• Không có kết quả nào được tìm thấy

Manual of Digital Earth

N/A
N/A
Nguyễn Gia Hào

Academic year: 2023

Chia sẻ "Manual of Digital Earth"

Copied!
296
0
0

Loading.... (view fulltext now)

Văn bản

Lü GN, Batty M, Strobl J et al (2019) Reflections and speculations on progress in geographic information systems (GIS): A geographic perspective. Zhang F, Hu MY, Che WT et al (2018) Framework for Virtual Cognitive Experimentation in Virtual Geographic Environments.

Transformation in Scale for Continuous Zooming

Continuous Zooming and Transformation in Scale

An Introduction

Continuous Zooming: Foundation of the Digital Earth

Transformation in Scale: Foundation of Continuous Zooming

Transformation in Scale: A Fundamental Issue in Disciplines Related to Digital Earth

1996), Scale in Remote Sensing and GIS edited by Quattrochi and Goodchild (1997), Scale Dependence and Scale Invariance in Hydrology edited by Sposito (1998), Modeling Scale in Geographical Information Science edited by Tate and Atkinson (2001), Scale and Geographic Inquiry: Nature , society and method edited by Sheppard and McMaster (2004), Generalization of Geographic Information: Cartographic Modeling and Applications edited by Mackaness et al. Researchers have also published authored research monographs, e.g. Algorithmic Foundation of Multi-Scale Spatial Representation by Li (2007) and Integrating Scale in Remote Sensing and GIS by Zhang et al.

Theories of Transformation in Scale

  • Transformation in Scale: Multiscale Versus Variable Scale
  • Transformations in Scale: Euclidean Versus Geographical Space
  • Theoretical Foundation for Transformation in Scale

That is, the transformation in scale in fractal geographic space is very different from that in Euclidean space. In a fractal geographic space, the level of complexity cannot be restored by scaling.

Fig. 8.2 A series of maps covering Hong Kong Polytechnic University at different scales (extracted from Google Maps)
Fig. 8.2 A series of maps covering Hong Kong Polytechnic University at different scales (extracted from Google Maps)

Models for Transformations in Scale

  • Data Models for Feature Representation
  • Space-Primary Hierarchical Models for Transformation in Scale
  • Feature-Primary Hierarchical Models for Transformation in Scale
  • Models of Transformation in Scale for Irregular Triangulation Networks
  • Models for Geometric Transformation of Map Data in Scale
  • Models for Transformation in Scale of 3D City Representations

Transformation model Large-scale Photo-reduce Small-scale. to move a line away from the position because it is too close to another feature). to remove the line because it is too small to be included). For the transformation to scale of 3D features, a set of models is listed in Table8.7, which is a summary of models proposed in the literature.

Fig. 8.8 Feature-primary and space-primary representations of spatial features: vector and raster models
Fig. 8.8 Feature-primary and space-primary representations of spatial features: vector and raster models

Mathematical Solutions for Transformations in Scale

  • Mathematical Solutions for Upscaling Raster Data
  • Mathematical Solutions for Downscaling Raster Data
  • Mathematical Solutions for Transformation (in Scale) of Point Set Data
  • Mathematical Solution for Transformation (in Scale) of Individual Lines
  • Mathematical Solutions for Transformation (in Scale) of Line Networks
  • Mathematical Solutions for Transformation of a Class of Area Features
  • Mathematical Solutions for Transformation (in Scale) of Spherical and 3D Features

Classification by K means is achieved by minimizing the sum of the squared error over all K clusters (i.e., the objective function) as follows: . where C¯kis is the mean of the clusterCk. In this section, two classical algorithms are described in detail, namely the Douglas-Peucker algorithm (Douglas and Peucker 1973) and the Li-Openshaw algorithm (Li and Openshaw1992).

Figure 8.24c-e show the results with different options, e.g., random selection and central pixel
Figure 8.24c-e show the results with different options, e.g., random selection and central pixel

Transformation in Scale: Final Remarks

However, the combined results are very irregular and the simplification of boundaries could be debated. Mathematical solutions for the transformation of spherical (e.g. Dutton1999) and 3D features (e.g. Anders2005) have also been investigated, although the literature is much smaller than that of map generalization.

Li ZL, Openshaw S (1992) Algorithms for objective generalization of line features based on the natural principle. Li ZL, Zhou Q (2012) Integration of linear and areal hierarchies for continuous multiscale representation of road networks.

Big Data and Cloud Computing

Introduction

Digital Earth data is of course big data due to the variety of data sources and huge amount of data. The increasing availability of big Earth data has provided unprecedented opportunities to understand the Earth in the Digital Earth context.

Table 9.1 Definition of the “9Vs” of big data
Table 9.1 Definition of the “9Vs” of big data

Big Data Sources

Remote sensing data is big data because of its large volume, variety, veracity and variability. Infrastructure data is big data because of its volume, velocity, variety, veracity, vulnerability, validity and volatility.

Big Data Analysis Methods

  • Data Preprocessing
  • Statistical Analysis
  • Nonstatistical Analysis

In Digital Earth, Yang (2011b, 2016) used association rules to mine the variables of Atlantic hurricanes from 1980 to 2003 and found a combination of factors related to the probability of rapid intensification, low vertical horizontal wind shear (SHRD=L), high humidity in the 850 –700 hPa (RHLO=H) and tropical cyclones in the intensifying phase (PD12=H). Linked data is structured data in which data sets are related to each other in a collection, which is useful for semantic queries and inference (Bizer et al. 2011).

Architecture for Big Data Analysis

  • Data Storage Layer
  • Data Query Layer
  • Data Processing Layer

For example, GEOSS is a cloud-based framework for global and cross-disciplinary sharing, discovery and access to Earth observation data (Nativi et al.2015). An in-memory data structure called the Resilient Distributed Dataset (RDD) manages datasets distributed in a Spark cluster (Zaharia et al.2012). GeoSpark provides operational tools for spatial big data processing based on Spark (Yu et al.2015).

Cloud Computing for Big Data

  • Cloud Computing and Other Related Computing Paradigms
  • Introduction to Cloud Computing
  • Cloud Computing to Support Big Data
  • EarthCube
  • Data Cube

The following section discusses the use of cloud computing to support big data management in Digital Earth. Three main categories of cloud computing services are (1) Infrastructure as a Service (IaaS), (2) Software as a Service (SaaS), and (3) Platform as a Service (PaaS). A sample of EarthCube efforts that take advantage of big data and cloud computing are introduced below.

Fig. 9.3 Cloud computing for big data analysis
Fig. 9.3 Cloud computing for big data analysis

Conclusion

Giuliani G, Lacroix P, Guigoz Y et al (2017) Putting GEOSS services into practice: a spatial data infrastructure (SDI) capacity building tool. Huffman GJ, Bolvin DT, Braithwaite D et al (2015) NASA global precipitation measurement (GPM) integrated multi-satellite queries for GPM (IMERG). Yang C, Yu M, Hu F et al (2017a) Using cloud computing to address major geospatial data challenges.

Artificial Intelligence

Introduction

In this chapter, we discuss artificial intelligence and machine learning techniques that have been used to manage and process geospatial data sets. We begin by discussing some traditional and statistical approaches in machine learning, and then introduce newer learning techniques applied to geospatial datasets. In this chapter, we describe some applications of these machine learning techniques to process the geospatial data sets that are the main content of Digital Earth.

Traditional and Statistical Machine Learning

  • Supervised Learning
    • Random Forest
    • Geographically Weighted Regression
    • Active Contours and Active Shapes
  • Unsupervised Learning
    • SKATER Algorithm
    • Autoencoders
  • Dimension Reduction
    • Evolutionary and Agent-Based Techniques

Using Mitchell's (1997) description, performance Q can be measured using a loss function where, for a given method and set of parameters, a location function, μ(x), is minimized (Hastie et al.2001) in Eq.10.2 . PCA has been used in various applications related to geospatial data representation and geospatial data analysis (Demšar et al.2013). GAs have been used in many applications in geospatial data analysis such as road detection (Jeon et al.2002) and satellite image segmentation (Mohanta and Binapani2011).

Fig. 10.1 Cartoon representation of a random forest classifier
Fig. 10.1 Cartoon representation of a random forest classifier

Deep Learning

  • Convolutional Networks
  • Recurrent Neural Networks
  • Variational Autoencoder
  • Generative Adversarial Networks (GANs)
  • Dictionary-Based Approaches
    • Dictionary Decomposition
    • Dictionary Optimization
  • Reinforcement Learning

Result (bottom left) and ground truth reference (bottom right) (courtesy of O. Arguda et al.). Conditional GANs have recently been used to automatically generate digital elevation models from user sketches (Guérin et al. 2017). According to the dictionary, signal decomposition consists of finding the best atom, ie. of an atom that increases the projection.

Fig. 10.10 Convolutional layers use fewer coefficients and are spatialized
Fig. 10.10 Convolutional layers use fewer coefficients and are spatialized

Discussion

  • Reproducibility
  • Ownership and Fairness
  • Accountability

Artificial intelligence and especially machine learning and deep learning have great potential to contribute to the generation, analysis and management of geospatial datasets. The digital earth should take advantage of such possibilities, as a placeholder to represent such data sets and a platform to analyze them. However, there are still issues in using them on Digital Earth platforms that need to be addressed.

Conclusion

Cootes TF, Taylor CJ, Cooper DH et al (1995) Active shape models-training and their application. Guérin É, Digne J, Galin É et al (2017) Example-based interactive terrain authorization with conditional generative adversarial networks. Lary DJ, Alavi AH, Gandomi AH et al (2016) Machine learning in geosciences and remote sensing.

Internet of Things

Introduction

In Section 11.2, we provide an overview of the most common definitions of the IoT, describe our working definitions in this chapter, and briefly discuss related work in the interaction between the IoT and the DE. In Section 11.3, we analyze the existing trade-off between both infrastructures in the context of the key high-level functions of DE. Then in section 11.4 you will find an overview of relevant case studies on different smart scenarios where the symbiosis of the Internet of Things and DE could lead to beneficial outcomes.

Definitions and status quo of the IoT

  • One Concept, Many Definitions
  • Our Definition
  • Early Works on the Interplay Between DE and the IoT
  • IoT Standards Initiatives from DE

Li and his colleagues investigated the impact of IoT on DE and analyzed the transition to Smart Earth (Li et al. 2014). The work of Van der Zee and Scholten (2014) highlighted the importance of location in the concept of IoT. The authors concluded that these technologies were already available for use in IoT and recommended their immediate use.

Fig. 11.1 Dimensions of the IoT (inspired in ITU-T 2012)
Fig. 11.1 Dimensions of the IoT (inspired in ITU-T 2012)

Interplay Between the IoT and DE

  • Discoverability, Acquisition and Communication of Spatial Information
  • Spatial Understanding of Objects and Their Relationships
  • Taking Informed Actions and Acting Over the Environment (ACT)

However, we need to interpret and contextualize these high-level features of DE from an IoT point of view. Of the two main capabilities of Things (see Section 11.2.2), the ability to observe and sense is a fundamental mechanism for providing observational data for DE. Acknowledging the blurriness of the boundary between the two infrastructures, we pay special attention to the trade-off between DE and the IoT in Figure 11-6, which shows how collaboration.

Fig. 11.6 IoT and DE workflow according to the higher cognitive functions in DE
Fig. 11.6 IoT and DE workflow according to the higher cognitive functions in DE

Case Studies on Smart Scenarios

Furthermore, van Setten et al. 2004) supported the COMPASS tourist mobile application with context-aware recommendations and route planning. The system has been extended with additional components such as a web-based client (Sawant et al. 2017). Regarding the delivery and reception of cultural heritage and cultural services, Chianese et al. 2017) proposed and tested a system that combines business.

Frictions and Synergies Between the IoT and DE

  • Discoverability, Acquisition and Communication of Spatial Information
  • Spatial Understanding of Objects and Their Relationships
  • Taking Informed Actions and Acting Over the Environment

One friction between DE and IoT relates to how geographic features are modeled. As mentioned above, DE needs to adapt to the opportunities that IoT devices can provide to enrich the capabilities of current IDIs. To conclude, Table 11.1 below summarizes the frictions and synergies between IoT and DE.

Conclusion and Outlook for the Future of the IoT in Support of DE

Præsenteret på 2018 IEEE 4th World Forum on Internet of Things (WF-IoT), IEEE, Singapore, pp. Jazayeri M A, Huang C Y, Liang S H (2012) TinySOS: Design and implementation of interoperable and tiny web service for the internet of ting. Van der Zee E, Scholten H (2014) Spatial Dimensions of Big Data: Application of Geographical Concepts and Spatial Technology to the Internet of Things.

Table 11.1 Detected frictions and synergies between the IoT and DE Discoverability,
Table 11.1 Detected frictions and synergies between the IoT and DE Discoverability,

Social Media and Social Awareness

  • Introduction: Electronic Footprints on Digital Earth
  • Multifaceted Implications of Social Media
  • Opportunities: Human Dynamics Prediction
    • Public Health
    • Emergency Response
    • Decision Making
    • Social Equity Promotion
  • Challenges: Fake Electronic Footprints
    • Rumors
    • Location Spoofing
    • Privacy Abuse
  • From Awareness to Action
    • Modeling the Geographies of Social Media
    • Detecting Location Spoofing Through Geographic Knowledge
    • Connecting Social Media with the Real World
  • Conclusion

Social media platforms can be used to limit the spread of pandemics and the fear associated with them. Any effort that ignores the importance of social media will question the effort. Ye X, Li S, Sharag-Eldin A et al (2017) Social media geography in public response to policy issues.

Digital Earth for Sustainable Development Goals

  • Fundamentals of Digital Earth for the Sustainable Development Goals
  • Information and Knowledge Relevant to National Implementation of the SDGs
    • How the SDGs Are Monitored and Reported
    • Information Needs for Implementation of the SDGs
  • State of the Art for the SDGs in DE
  • Case Study of Australia: Operationalizing

Section 13.2 identifies countries' information needs for implementing the SDGs, including for the SDG Global Indicator Framework (GIF). Section 13.3 summarizes the findings of recent research and practice on the use of Digital Earth (including Earth observation1 and social sensing) in support of the SDGs. Specific to the EO community are the challenges to consistently and systematically transform satellite and other remote sensing data into valuable global information layers to support effective implementation of the SDGs.

Fig. 13.1 National implementation of the SDGs requires evidence-based approaches for monitoring and reporting
Fig. 13.1 National implementation of the SDGs requires evidence-based approaches for monitoring and reporting
  • DEA to Map Land Cover and Dynamics Over Time
  • DEA in Support of SDG Indicator 15.3.1
  • Digital Earth in Support of SDG 17: Strengthen Means of Implementation
  • The Way Forward: Partnerships to Strengthen DEA in Support of the SDGs
  • Big Earth Data for the SDG: Prospects
    • R&D and Technology
    • Normativity, Governance and Institutional Arrangements
    • Science-Policy Interface
  • Conclusion

Giuliani G, Chatenoux B, De Bono A et al (2017) Building an earth observation data cube: lessons learned from the Swiss Data Cube (SDC) on generating analysis-ready data (ARD). . Lymburner L, Bunting P, Lucas R et al (2019) Mapping multi-decadal mangrove dynamics of the Australian coastline. Paganini M, Petiteville I, Ward S et al (2018) Satellite earth observations in support of sustainable development goals.

Fig. 13.5 Examples of data inputs for the application of the FAO LCCS level 3 within Digital Earth Australia used to produce standardized land cover maps at 25 m resolution
Fig. 13.5 Examples of data inputs for the application of the FAO LCCS level 3 within Digital Earth Australia used to produce standardized land cover maps at 25 m resolution

Digital Earth for Climate Change Research

Hình ảnh

Fig. 8.1 A series of images covering HK Polytechnic University at different scales (from Google Maps)
Fig. 8.2 A series of maps covering Hong Kong Polytechnic University at different scales (extracted from Google Maps)
Fig. 8.6 The natural principle: spatial variations within a smallest visible size (SVS) to be neglected (Li 2007)
Fig. 8.8 Feature-primary and space-primary representations of spatial features: vector and raster models
+7

Tài liệu tham khảo

Tài liệu liên quan