The field of surveying is one of the most highly automated fields in civil engineering. If you study the use of application software within an engineering firm, you will often find that only 10 to 15 percent of the software's capabilities are used throughout the company. But, if you break usage down by department, you will undoubtedly hear of significantly higher numbers for surveying, perhaps even in the 70 to 90 percent range. Surveyors use computers for field data collection (whether aerial or survey field crew), reduction, research, computations and preparing for stakeout. Perhaps civil designers can take some advice from these automation pioneers: "Work in Three Dimensions."

## 3-D Pioneers

Surveyors have been working in 3-D longer than many civil engineering professionals. They routinely collect, process and compute site data in three dimensions. Construction staking is also performed in 3-D because points can be loaded directly to the field recorders from the CAD system. So, why are site grading computations in civil engineering still performed in two dimensions? This requires surveyors to re-analyze this data and computationally extract what they need to stake out the proposed construction.

If other design professionals work and portray their work in 3-D, why is it that civil engineering lags so far behind? Mechanical engineers, aeronautical engineers and architects largely work in at least three dimensions. If civil designers performed site design and developed CAD elements using three dimensions, surveyors would have a better chance of understanding the design, reviewing it and extracting survey stakeout information directly from the design.

Figure 1. Example of Mass Points and Breaklines.

## Mass Points and Breaklines

When surveyors use electronic field recorders to collect existing conditions for a project, they collect breaklines and augment them with spot shots. Spot shots and breaklines consist of northings, eastings and elevational data. Breaklines are three-dimensional linestrings (or polylines) where each vertex may have a different elevation. Breaklines represent grade breaks on the ground; examples include the edge of pavement, the top of a bank or the crown of a paved road. In Figure 1, note that the breaklines portray features such as shoulder, edge of pavement, crowned pavement, sidewalks, creeks, and curb and gutter sections.

Figure 2. A TIN from contours.
Photogrammetrists collect essentially the same data but cover larger areas. An aerial surveyor who wants to provide a highly CAD-conducive dataset will capture not contours, but mass points and breaklines. As a case in point, the file size for the contours of a 750-acre site with significant terrain can be as high as 20 Mb. The file size for the same site captured in 3-D using mass points and breakline data can be as low as 0.75 Mb.

In addition to the storage resources needed to handle such large CAD files, the quality of the data available for designers is enhanced if breakline data is used. For instance, a designer might need to know the elevations along the edge of pavement. If traditional contour data represents that surface, then a straight-line interpolation would need to be computed to estimate the elevations. Although this is a reasonable assumption, it could produce a somewhat incorrect answer.

A major detraction against the use of contours is that the elevation is known only along the contour. The ground activity between contours is open to educated speculation. The strength of breakline data is that an infinite number of elevations exist throughout the length of the object, and this is actually a truer representation of what the surveyor shot.

Another factor is that the system resources allocated for the processing of contour data is much higher than for processing breaklines and spot shots. When a digital surface is prepared, it is done based on a triangulated irregular network or TIN. Each triangle within the TIN connects the closest three points on the surface, whether they are spot shots or breakline vertices. Consider that contour vertices are quite close together, many times within 2' to 5' of each other. That means triangles will be generated every 2' to 5' along the contour. If breakline data were collected every 25', then the size of the TIN would be greatly reduced when compared to the contour-produced TIN. Note that the TIN produces virtually all of the surface computations for the project. In other words, profiles, cross-sections, storm structures, waterline depths, etc., are usually extracted directly from TIN information. In Figure 2, the TIN shows 10 contours that produce 1,986 triangles representing the surface. The high density of this TIN does nothing but consume system resources every time it is accessed for information.

So, not only are designers carrying huge contour files with them for the duration of the project, they are also carrying huge TINs with them as well. Therefore, when beginning a new project, make sure that large sites make use of mass points and breaklines. This is natural for field surveyors since they capture points and breaklines as a matter of routine. But, project managers should specify that aerial surveyors produce a deliverable that contains mass points and breaklines in addition to the traditional contours. The field-supplemented data can be readily integrated into the overall aerial topography using standard CAD modification commands.

When the project data moves into the design arena, the designer should have access to the 3-D data for analysis purposes. The surveyor should produce a correct TIN and related contours, but there is no harm in allowing engineers to check the source 3-D data when needed.

As design progresses, many of the popular road design programs can output the proposed roadway surface in 3-D at user-specified intervals. They will also create 3-D meshes for quality assurance and rendering purposes. An increasing number of road designers are performing these tasks. More often than not, this is where the use of three dimensions usually stops in our industry. Most designers then spend their time drawing contours and computing spot shots to reflect the proposed grading.

A designer thinks in three dimensions using elevations, slopes and grades. Then they convert that thinking into two-dimensional contours and potential grading information if lost. If the engineer wants layout for the flowline for a curb section, he or she typically think as follows: "Begin at an elevation of 101.44 and drop the flowline at two percent for 250'." Using CAD tools over a decade old, we can instruct the software to use these same parameters and let the software compute where contours are located. When changes occur, a technician can use CAD tools to make modifications. The curb section can be raised or lowered en masse, and line segments can be easily adjusted.

## Parametric Site Design

Two of the most popular CAD software manufacturers, Autodesk with their Land Development Desktop (San Rafael, Calif.) and Bentley/GEOPAK (Exton, Pa.), have been moving toward parametric technology for civil site design. This embeds design data into CAD elements and exhibits itself when moving those objects, raising/lowering objects and performing earthworks. For instance, a stormwater retention pond can have its outside berm tied-out to natural ground with a sideslope parameter of 3:1. When the pond is moved using standard CAD tools, the sideslope tie-outs re-grade themselves as the natural ground under the pond changes.

The real power of a 3-D surface is the capability of a designer to identify surface elevations anywhere, automatically. Slopes of flowlines, elevations of the edge of pavements or sidewalks and other related information can be achieved at the touch of a cursor. Modifications can be accomplished easily and effectively, and the resulting 3-D models can be easily exported to programs where they can be rendered and animated.

Figure 3. Proposed breaklines.

## Proven 3-D Projects

Planar surfaces such as an asphalt parking, sidewalks, building foundations and walls are built using 3-D faces in the CAD system. Rolling terrain for features such as landscaping berms were represented by contour data.

Once the surface was developed, the software interpolated the location of contours and labeled them. Changes to the surface could be made on an as-needed basis and the contours could be re-imported for construction plan purposes.

After the final design was complete, the design loop closed, and the proposed data was sent to the survey department for stakeout and plat preparation. How easy is it for the surveyor to extract surface values such as spot elevations and slopes? If a 2-D design is made available to surveyors, he or she must calculate grades based on spot shots. If a 3-D design is made available, then much of this data can be extracted directly from the design. Slopes and spot shots can be extracted at an infinite number of locations throughout the project.

In Figure 3 on page 87, note the breaklines placed in 3-D to grade a commercial shopping plaza.

Figure 4. Resulting 3-D surface.
In Figure 4, engineering software develops 3-D faces from the breaklines and point data.

Figure 5. Rendered 3-D surface from 3-D faces.
Figure 5 shows the rendered 3-D surface of the proposed grading for a shopping plaza. It is directly renderable and any surface information is available at the touch of a CAD command. This level of quality control is certainly one of the benefits of working in 3-D. Truck loading docks, parking lot islands and even handicap ramps are reflected in the design work.

Figure 6. Cross-sections can be computed instantaneously.
Cross sections can be computed instantly from the surface data because of this (see Figure 6).

Civil engineers can learn from their surveying compatriots. Surveyors work in three dimensions from field collection through stakeout. If civil designers designed in 3-D, quality assurance would benefit and surveyors would be able to expedite reviews of the projects. Stakeout tasks would flow more smoothly between the project database and electronic field recorders.