This concept paper on the automatic acquisition of 3D city models was written by Venugopalan Nair who is the Senior Manager of Geo Spatial Services at SBL.
Urban areas are replete with a variety of geospatial information. Some of this spatial information is virtual and some contains very complex physical information. Virtual information includes administrative boundaries such as city boundaries, electoral districts, and other jurisdiction boundaries of various governmental departments. Physical information includes transportation networks, communication networks, utility networks, land use land cover features, and hydrographic networks, to name just a few. Collecting this information is essential to local government function; all queries related to a city or urban management system will definitely relate to some location or spatial reference. Though two dimensional maps are available and common, 3D city viewing is rapidly gaining acceptance and application especially in light of smart city initiatives. Application of 3D city models range from citing of tower locations of cell phone companies to micro climate, environmental, air pollution, architectural designs and urban heat wave analysis.
3D city models are digital three dimensional models of urban areas with all its features. These features include aforesaid virtual features as well as other physical furnishings such as buildings, land marks, vegetation, infrastructure landscapes and terrain. These features can be represented in spatial data base in the form of two dimensional as well as three dimensional features class. 3D city models are visually integrating various geo spatial elements to real world scenarios with all its complexities. Creation of automatic 3D city models involves preparation of data base based on the required level of details, storage of the same in a data base, construction and detailing of city models and visualization of the same through various application platforms.
Some of the key applications of 3D city models are listed below.
Master planning applications: 3D city models are an apt tool for the town planners. 3D city models will help planners from conceptual development stage to the completion and long term environmental impact assessment stage. The benefit of 3D city models in two planning is that it will reduce planning errors and hence monitory loss, it provides better visualization capabilities and hence monitoring is very easy. The most import part played by 3D city models in such application is that it will greatly reduced project delays. It can also add value as a depicter which can monitor illegal vertical growth of the city.
Spatial Database Infrastructure Applications: 3D city models are part of the spatial data base infrastructure of the region. Irrespective of the initial cost of preparing such data base, in the long term it will be the robust data base service in the back ground of all developmental activities. This will be a base data for further spatial data simulation and modeling. Site suitability analysis of roof top for solar panel installations, thermal emission analysis of buildings and urban fabric are some of such analysis.
Disaster Management applications: 3D city models are highly helpful in disaster management in terms of providing near realistic effected of the disaster. It can simulate fire, flood, bomb explosions etc and after effects on the same in urban infrastructure in place to a great precision. Such simulations will help administrators and planners in managing their rescue operations and resource mobilization.
Navigational applications: 3D building models and terrains will give a realistic display of the area for people navigating and crossing through the area. It is gaining widespread acceptance than simply looking at the locations. It can be act as a tool for intelligent transportation system.
Utility design applications: 3D city models will play a key role in designs of sewerage channels, railway line routing, highway routing, power transmission lines routing and other geo designs.
Archaeological applications: Historical buildings and cultural heritages can be modeled to a highest level of detail (LoD4 ) in 3D city models so that its intricacies will be preserved and will help in maintenance and repairing.
Decision support system applications: 3D city models can be aptly used by planners and managers to take informative decisions using interactive models.
Property management applications: 3D city models can be effectively used for property management and also for dispute settling tool.
Entertainment applications: 3D city models can be back ground and give virtual reality for gaming industry.
There are several kind of input used for preparation of 3D city models. These are ranging from very high resolution stereo aerial photographs to high precision LiDAR data to high resolution satellite images. The kind of input used and level of details (LoD) extracted are solely based on the requirement of 3D city model. Buildings are most important items in a 3D city Model. The LoD is mainly for the detailing of the buildings. There are five levels of standard LoDs and these are
3D city models have many kinds of data. These include but not limited to GIS data, CAD data, BIM data, and some of the non spatial data as attributes. GIS data is very much required for any 3D city model. Base maps, digital elevation models, digital terrain models, land use maps, and administrative boundary maps, road network data etc. are the GIS component of the 3D city model. CAD data play a key role in 3D city models. Most of the buildings with required level of details as mentioned above can be designed in a CAD environment. Other infrastructure such as railway network, road network and other utilities can be modeled in CAD environment. The outcome in the form of wire frames or surface models will add flavor to the 3d city models. BIM data will provide with one step ahead information of CAD data. The specifications of the buildings in the form of attributes can also be incorporated in CityGML data model.
3D city models comprise a variety of data sets, in a variety of formats. So there cannot be a uniform data base schema for the same. The data base schema will be depending on the requirement and application of the models. The component GIS data sets can be raster rata sets as well as vector data sets. 2D raster data sets can be in .geotiff, .img, .jpg like format based on the rendering capacity of the hardware as well as software and internet capability. 3D models again in ASCII, tiff or obj format where common CAD, GIS and graphic software can render and display. Reliable data base management is a prerequisite for complex analysis and result derivation from 3D city models. So a relational data base is very much required for running and utilizing any 3D city models. In this relational 3D city models, hierarchically structured, multi scale data can be stored, retrieved and analyzed. For Updation of the 3D city model data based is also possible only through a relational data base. Also the relational data base should able to hold multiple formats of the data and its rendering. Open geospatial consortium standards and norms can be followed in 3D city models creations as well. Such data bases are referred as city GML
One of the major challenges in automatic acquisition of 3D city models is its visualization and semantic rendering. Visualization is a challenge because component of a 3D city model will be mostly from different data sources such as geo data, GIS data, BIM, CAD etc. So a common frame work schema based on the requirement of the model need to be established beforehand through a relational data base. Visualization and rendering are based on the core functionality required through interactive application suites. Typically 3D city model rendering should be fast, scalable, and cost effective. In case of 3D geometry and 3D texture display a specialized algorithms need to be established. In general rendering can be service based or maps based. City GML will provide not only the visual 3D city models but attributes can also be viewed.
Service based rendering: This kind of rendering will based on data base architectures which runs on client applications which is external in nature. In this case all 3D portrayals of computer graphics and other GIS data sets will be the responsibility of the external application. This can be Open Geospatial Consortium (OGC) compatible. Services based rendering can be of Web 3D service or Web View Service. The former will be based on the client application which is converted the scene graphics and other geo data in 3D models and later is through server based services. In this case the server will generate 3D views which are uploaded on requested to screen and streamed based on the scale of rendering. Here client applications will generate 3D models and save in the server and rendering is not directly done by application.
Map based rendering: Smart 3D city models are following map based approach and render 3D visualization to any kind of platforms. In this case the 3D city model is arranged to a series of small tiles which enable automatic 3D rendering. Different levels of details of the model and pre arranged in a server which any number of concurrent clients can access. Map based rendering is more simplified in terms of data transfer and complexity of the 3D data models.
Buildings are the major and complex component of any of the city model data base. Building information model can be from simplest extruding footprints (LoD1) to very complicated architecture with interior designs (LoD5). Automatic extraction of buildings is possible from 3d laser point clouds and photogrammetric approaches. But through this kind of approaches only LoD1 and LoD2 is possible. For higher levels of details Digital terrain models and digital surface models are required. Operator interference and statistical approaches are required for higher level of details of the BIM.
The size of Mobile Mapping Market was anticipated to be $14 billion in 2020 and is predicted to reach $40…
Esri has unveiled a new massive open online course (MOOC) entitled, "Spatial Data Science: The New Frontier in Analytics."