Elsevier

Computers & Geosciences

Volume 89, April 2016, Pages 12-20
Computers & Geosciences

Research paper
An approach to 3D model fusion in GIS systems and its application in a future ECDIS

https://doi.org/10.1016/j.cageo.2016.01.008Get rights and content

Highlights

  • We propose a precise and efficient approach to fuse 3D models in GIS systems.

  • We implement our fusion approach in a future 3D ECDIS.

  • This approach solves the primary problem for the use of 3D models in the local system.

  • This approach can improve the application of 3D models in GIS.

Abstract

Three-dimensional (3D) computer graphics technology is widely used in various areas and causes profound changes. As an information carrier, 3D models are becoming increasingly important. The use of 3D models greatly helps to improve the cartographic expression and design. 3D models are more visually efficient, quicker and easier to understand and they can express more detailed geographical information. However, it is hard to efficiently and precisely fuse 3D models in local systems.

The purpose of this study is to propose an automatic and precise approach to fuse 3D models in geographic information systems (GIS). It is the basic premise for subsequent uses of 3D models in local systems, such as attribute searching, spatial analysis, and so on. The basic steps of our research are: (1) pose adjustment by principal component analysis (PCA); (2) silhouette extraction by simple mesh silhouette extraction and silhouette merger; (3) size adjustment; (4) position matching. Finally, we implement the above methods in our system Automotive Intelligent Chart (AIC) 3D Electronic Chart Display and Information Systems (ECDIS). The fusion approach we propose is a common method and each calculation step is carefully designed. This approach solves the problem of cross-platform model fusion. 3D models can be from any source. They may be stored in the local cache or retrieved from Internet, or may be manually created by different tools or automatically generated by different programs. The system can be any kind of 3D GIS system.

Introduction

3D models are widely used in GIS, medical field, mechanical field, movie, game, and so on. They have become the fourth data type of multimedia after sound, image, and video (Cui and Shi, 2004). As an information carrier, 3D models are more visually efficient, quicker and easier to understand and can express more detailed information than traditional two-dimensional (2D) graphics. They can dramatically improve users' awareness of the environment and cognition of the objects they describe. GIS researchers start to use 3D models to describe spatial objects such as buildings, terrains, and so on. This significantly improves the cartographic expression and design and greatly promotes the development of GIS applications.

With the development of 3D information systems such as digital earth, digital city, and so on, there is an increasing demand for 3D models. Development and maturation of 3D modeling tools make the modeling process easier and easier. Tools such as 3DS Max, SketchUp, Building Information Modeling (BIM) tools, and so on, help users conveniently create 3D models. Increasingly advanced Internet technology provides the conditions necessary for people to conveniently share and search 3D models. This promotes the development of 3D models. All these have led to an explosion in the number of available 3D models on the Internet and in domain-specific databases. And these have resulted in the rise of a 3D model retrieval platform for public usages (Funkhouser et al., 2003, Chen et al., 2003, Zhang, 2005, Tangelder and Veltkamp, 2008). A number of 3D model data-sharing databases and data centers have been established, such as 3D Warehouse, 3D Source, and so on. Moving Picture Experts Group (MPEG) researched multimedia content description interface standard (MPEG-7) to provide the standardization of content-based description and retrieval specifications of 3D models (Sikora, 2001). And MPEG-7 made the web as searchable for multimedia content as it is searchable for text. This further promotes the application of 3D models.

Many researchers in the GIS area have studied the application of 3D models to provide 3D systems, aimed to reduce the cognitive load and provide more rich and intuitive spatial analyses. Coors (2003) presented a data model for the search of 3D models and studied the visualization of 3D models in 3D GIS. Wu (2004) presented a universal 3D model, generalized tri-prism for 3D geoscience modeling system and real-3D GIS. Yu et al. (2004) introduced a novel approach to the mesh editing of 3D models with the Poisson equation as the theoretical foundation. Apel (2006) proposed an integrated data model for geological observation data and geomodels and presented an approach for querying observation data and 3D geomodels based on their 3D spatial and geological properties. Gruen (2008) presented the urgent need for the generation, administration, analysis, representation and manipulation of 3D models in digital earth. Isoda et al. (2009) explored an approach for creating large-scale urban 3D models using GIS data. Thompson et al. (2011) analyzed the emerging issues for the creation of an information rich and geometrically accurate 3D model for two adjacent cities with the conclusion that 3D models could offer much greater accuracy and efficiency for the planning process. Xu and Coors (2012) combined system dynamics model, GIS and 3D visualization to assess the urban residential development. Mali et al. (2012) used 3D city models to give the precise information support to management policies and future planning of urban resources. Jakubiec and Reinhart (2013) used detailed geometric urban massing models and Daysim-based hourly irradiation simulations to predict city-wide electricity gains from photovoltaic panels.

In the aspect of ECDIS, since Ford (2002) first introduced 3D navigational charts, many researchers have studied 3D charts to provide chart display systems, aimed to reduce the amount of marine accidents caused by fatigue, mental overload and limited awareness of the navigational situation. Gold et al. (2005) presented a prototype of an interactive 3D “pilot-book” with 3D model (map) derived from Electronic Navigational Chart (ENC). Porathe (2006) proved that compared with traditional 2D charts, 3D charts were more efficient and could dramatically reduce the number of human errors. Ternes et al. (2008) proposed a type of 3D visualization system that worked with manually prepared 3D models, for the support of navigation during hydrographic surveys. Ray et al. (2011) introduced a 3D chart to facilitate the understanding of maritime behaviors and patterns at sea. Liu et al. (2014) proposed a type of future ECDIS, where 3D models in this system are fused with 2D charts to describe entities in the navigation environment such as buoys, buildings, wharves, bridges, etc.

There is a huge potential of applications in reach, if we can take advantage of the 3D model in overwater and overland GIS applications. It should be noted that, in all applications, we need to solve a problem first: fuse the 3D model into our own system or software. Since systems vary widely and there are so many available 3D models on the Internet and in databases, this problem will be more complicated. However, most of the researches listed above only focused on the construction and visualization of 3D models. Few researchers studied how to automatically, precisely and efficiently fuse 3D models in GIS systems, especially in ECDIS.

In this paper, we propose an automatic and precise approach to fuse 3D models in GIS systems based on 2D vectors. 2D vectors can be extracted either from maps and charts or from remote sensing images. 3D models from different sources may be in any pose, size, and position. First, we give some basic concepts and describe our system AIC 3D ECDIS. Then we introduce the fusion approach. The approach is mainly divided into four parts: pose adjustment, silhouette extraction, size adjustment and position matching. At last, we implement our method in our system AIC 3D ECDIS to provide a more precise and efficient 3D depiction for a future ECDIS. Also the application shows the effectiveness of our approach. The automatic and precise fusion of 3D models in GIS systems is studied.

Section snippets

Basic concepts and AIC 3D ECDIS

In order to give a detailed introduction to our fusion approach, this section first introduces some basic concepts. Section 2.1 gives the concept of model representation; Section 2.2 gives the definition of simple mesh, which is an important concept in our fusion approach; Section 2.3 introduces our system AIC 3D ECDIS; Section 2.4 describes exiting problems in the 3D model fusion.

Fusion approach

In order to fuse 3D models from different sources in local systems, we need to adjust the pose and size of the model and then set the position. In this paper, we propose a precise and automatic method to deal with the 3D model fusion in GIS systems. This approach solves the problem of cross-platform model fusion. Firstly, use PCA to adjust the pose of the model. The model can then be projected, adjusted and fused. Secondly, project the selected meshes of the model onto the specified plane and

Application

We implement our fusion approach in our system AIC 3D ECDIS to provide a more precise and efficient 3D depiction for a future ECDIS. Also, the application shows the effectiveness of our approach.

There are two ways of using our fusion approach: either to implement the approach on the server side and publish the model and transformation parameters to the client side, or to implement the approach in the client program.

In order to simplify the client’s workload, we choose the first way. The model

Conclusion

This paper provides a precise and efficient approach for the fusion of 3D models in GIS systems and implements the approach in our system AIC 3D ECDIS. This approach solves the primary problem for the use of 3D models in the local system.

However, the 3D model may not be normatively constructed. The model may contain useless or unnecessary points and lines. And the proportional relationship between different parts of the model may be incorrect. In these cases, the model should be modified before

Acknowledgments

We would like to thank the reviewers and the editor for their helpful and constructive suggestions for improving the paper.

References (29)

  • C. Gold et al.

    3D geographic visualization: the marine GIS

  • A. Gruen

    Reality-based generation of virtual environments for digital earth

    Int. J. Digit. Earth

    (2008)
  • Y. Isoda et al.

    Reconstruction of Kyoto of the Edo era based on arts and historical documents: 3D urban model based on historical GIS data

    Int. J. Humanit. Arts Comput.

    (2009)
  • M.M. Kazhdan

    Shape representations and algorithms for 3D model retrieval

    (2004)
  • Cited by (12)

    • Achieving smart sustainable cities with GeoICT support: The Saudi evolving smart cities

      2017, Cities
      Citation Excerpt :

      Dangermond (2009) and Abukhater and Walker (2010) demonstrated how GeoDesign could be used to foster sustainable design process. Three-dimensional city modelling can assist planning through enhanced visualization (Dorninger & Pfeifer, 2008; Liu, Zhao, & Pan, 2016). A 3D city model may serve as an input into other urban models and can be used in telecommunication and disaster management.

    • 3-D wavelet-based fusion approach for comprehensively analyzing multiple physical-property voxel models inverted from potential-field data

      2017, Journal of Applied Geophysics
      Citation Excerpt :

      The techniques for data fusion are in general divided into two main categories: spatial-domain fusion (such as principal component analysis, PCA) and transform-based fusion (such as discrete wavelet transform, DWT). Currently, these techniques are applied widely in the fields of computer science, engineering and medicine (Li, 2006; Gökberk et al., 2008; Liu et al., 2016). However, most of the applications of fusion in the geophysical literature are based on the 2-D fusion techniques which were used only on 2-D data, such as gravity and magnetic anomalies, rather than on 3-D dataset.

    View all citing articles on Scopus
    View full text