Example: Classifiers of the Philadelphia Reality Model

In 2015, a reality model of Philadelphia was created to support the Pope’s visit to Philadelphia. In order to demonstrate Reality Model Classification, we will link this model with GIS data and explore the value of this combination. Like most municipalities, the GIS data for Philadelphia is readily available. The website www.opendataphilly.org provides over 300 different datasets. Let's start by downloading the building footprint data in SHP format.

The SHP data can be opened directly in MicroStation and saved to DGN format for convenience. We also directly open the reality model in the “.3mx” format created by the ContextCapture software and save that to DGN.

Both of the DGN files are geographically located. However, they may not use the same geographic coordinate systems. Fortunately, MicroStation can reproject the GIS data to match the coordinate system of the reality model. It is not possible, however, to reproject reality model data efficiently between geographic coordinate systems (the sheer volume of data makes this operation impractical). Therefore, it is important to use the reality model as the master and attach as a reference (and potentially reproject) the design data to the reality model geographic coordinate system.

After attaching the building footprint data, you can view the model from below and observe that the footprint data matches the reality model very well, confirming that both the reality model and the GIS data were geospatially accurate.

The next step is to attach the building footprint reference as a classifier to the reality model. You can use the Attach Reality Mesh tool for this. In the tool settings, select Type as Model so that all of the building footprints in the referenced model are part of the classification and set the Name as Building Footprints. Set Margin to 1.0 (meters), which designates that each classification volume will be one meter larger than the building footprint. A single element within the footprint reference attachment is selected to designate that this is the desired model for the classification.

Once you attach the classification, it will appear in the lower list box of the Reality Mesh Attachment dialog and is set to be the Active Classifier for the reality model attachment.

To interactively view the result of the newly attached classification you can use the Select Reality Mesh Classifier tool and move the cursor over the reality model. When the reality mesh below the cursor is within a classification boundary, the classified volume is flashed in the highlight color. Picking the reality mesh classification will add the classification boundary object to the current selection set and the Properties dialog can be used to view the properties of the object.

Now that the classification object is selected, you can use the Classifier settings to control its displayed. The Inside and Outside settings control how the reality mesh data inside or outside the boundaries of the active classification boundaries are displayed. The Selected setting controls how the reality mesh for selected classification boundaries is depicted. By turning both Inside and Outside Off and changing Selected to On, you can isolate to display only the reality mesh for the selected classification boundary (see below). Note that the Select Reality Mesh Classifier tool continues to function with the Inside and Outside settings Off, so in this mode it is possible to continue to interactively select from the (now invisible) reality mesh.

With different settings of the Inside and Outside, you can view the classification data in interesting ways. With Inside: On and Outside: Off, you can display only the buildings.

By setting the Outside: Dimmed, you can see the buildings against a dimmed background of the unclassified reality mesh. Note that in this case the dimmed building in the center is new and missing from the building footprint data.

Many different classifications can be attached to a single reality model. This provides many different ways of viewing the same reality model. By attaching a Classifier of the commercial corridors you view the same model as a set of zones rather than individual buildings. Note that in this case, a negative margin distance is used to inset the commercial zones slightly to produce a small Outside zone at their boundaries.

In most cases, it is preferable to use Classifiers to control how the reality model is viewed without changing the reality model attachment. It is, however, possible to use the reality mesh clipping tools to permanently clip the model at classification boundaries. For example, in the zoning example above, it is possible to create separate attachments for the commercial zones by selecting the zones (<Ctrl+Click> will allow selection of more than one classification boundary) and using the Split Reality Mesh tool to create separate attachments for each zone and a single attachment for areas outside the selected boundaries.

Point and Linear classifiers

2D Linear geometry is important in GIS, Civil and other disciplines as it is often used for roads, railroad tracks, waterways or boundaries. Point data is frequently used to represent locations of interest or vertical assets such as street signs or telephone poles. Although this geometry alone does not enclose an area that can be projected to produce a classification volume, it may be combined with the Margin setting to produce classification volumes that enclose and surround the geometry.

Example: Philadelphia with Arterial Streets Classifier

In the example in the first image below, the streets are used as classifiers for the Philadelphia Reality model. At first glance, it appears that there are streets that are incorrectly placed through buildings. On closer inspection (see second image), you see that these are roads that actually go through the structures – an anomaly that would be difficult to discern without 3D reality model.

The example below illustrates the use of a point classifier representing the location of street poles in the Philadelphia reality model.

Volumetric Classifiers

In the previous examples, 2D GIS data has been used to classify the reality meshes. In this case, the 3D geometry is projected to classify all reality model geometry above (or below). It is also possible to use 3D geometry to more precisely enclose reality mesh volumes. Currently, volumes can be specified by slabs, cylinders and non-parametric extrusions. If other 3D geometry is used, then the classified volume is determined by the range of the element.

In the example below, a classification of the individual floors of a building rather than the entire building is illustrated. A series of slabs provide the classifiers. Properties of the floors (occupants, rental rates, and so on) can be accessed by interactively selecting the reality model to access these classifiers.

Activating the “Tower Floor” classifier and Selecting Dimmed for the Outside display mode highlights the building of interest. Using the Select Reality Mesh Classifier tool, you can select the individual tower floors.

Example: Reality Mesh Classification with Explorer's Advanced Search

The reality mesh classification uses the standard MicroStation Selection Set system. It is possible to use the Advanced Search feature of the Explorer to search for Classification elements and to reflect the search results on the reality mesh. This is illustrated using a Reality Mesh created for the city of Coatesville, Pennsylvania. A Classifier with parcel data is attached to the reality mesh and a search on all parcels where the owner is "CITY OF COATESVILLE" is performed. The city owned portions of the reality model are displayed below.