Point-cloud software offers versatile solutions.

BY OMAR-PIERRE SOUBRA

Today, 3-D laser scanning technology is able to deliver robust data for civil engineering projects that are under construction or are complete. In addition to the spatial data expected from a survey, laser scanners are providing information such as true color (giving the impression of seeing a picture of the project in 3-D), returned laser intensity (which is the amount of light captured by the instrument, helping to depict the scene), and project management data (including location, date and time information, technicians' names, weather conditions, and scanner position). However, most of the software traditionally used to process surveying data for civil engineering projects is not designed to handle such huge amounts of data.

Therefore, the use of point-cloud processing software, which converts the data generated by 3-D laser scans into more useable formats for civil engineers, is necessary. Conversion can be into a full 3-D CAD model, a volume calculation, an inspection report, or even into standard 2-D drawings. Regardless of the deliverable, keep in mind that the original information - virtually a digital copy of a project - always will be available for archiving, future design, or, if necessary, forensic analysis.

The point-cloud The 3-D laser scanners used in civil engineering applications measure a project's current status very quickly. Typically, 5,000 single points are measured per second. And you can collect measurements every tenth of an inch for objects that are 300 feet away from the scanner. Therefore, the database of a bridge or a road scan can grow rapidly to include more than tens of millions of measurements, and standard CAD packages are unable to display the information.

For each point scanned, the X, Y, Z data (north, east, and elevation); the color of the point measured; and the returned intensity of the laser are collected, and together compose the point-cloud.

To collect the complete information about a project, a laser scanner may need to be situated in various locations to shoot from different angles. After the scanning is complete, the data collected from the various locations is stitched together.

Next, the point-cloud is divided into subsets of interest. This step has two purposes: It enables one to work with only the needed parts of a project, and it allows several people to work simultaneously on the same project, increasing productivity.

Deliverables 2-D drawings from 3-D scanner data - In the field, 2-D drawings remain the primary information used by contractors.

Two approaches are available to create these drawings from 3-D laser scanning data.

The first approach is to work solely with point-cloud data processing software, which can handle enormous datasets. This type of software automatically “tracks” (converts the point-cloud into 2-D drawings) natural features in the dataset such as corners, plans, and discontinuities. Some vendors offer advanced capabilities that enable automatic extraction and tracking of even complex shapes in the point-cloud such as rail tracks, sidewalks, street profiles, and road median barriers (virtually any linear object).

The second approach to creating 2-D drawings is to import low-resolution point-cloud data into standard CAD packages such as AutoCAD or MicroStation. This comprehensive information is sampled from the original point-cloud and usually contains a few hundred to less than 10,000 points. Alternately, ortho-rectified snapshots of the point-cloud (also know as orthoprojection) can be imported into CAD. Once the data is imported, the CAD operator can draw directly on it, using standard CAD tools.

3-D models - 3-D models can be created in two ways as well. For complex shapes, a mesh (which is a triangulated surface, similar to a digital terrain model) can be computed automatically by the pointcloud processing software. For simpler 3-D shapes, such as planes, rounded shapes, or cylinders, the operator can convert the point-cloud into solid 3-D shapes. These solid shapes usually are imported into simulation software, such as lighting/ illumination simulation programs or animation software, which can be used to test the impact of a project on the natural environment. This process is more time consuming than the automatic mesh creation, but vendors are improving it by creating more automatic fitting tools. These tools recognize the shape of a designated area in the point-cloud and fit the appropriate geometry to it. Moreover, for increased productivity, the 3-D shapes to be used can be linked or imported from an existing database that was built for other software.

Inspection - As-built inspection is probably among the most popular applications for 3-D scanning data, as it enables comparison of what has been built with original designs. This procedure is achieved in point-cloud processing software by importing the original 2-D or 3-D design (in a CAD file format) into the pointcloud processing software and asking for a comparison between the CAD file and the measured points. The software provides graphic visualization of gaps and deformations; in addition, advanced tools enable operators to export complete inspection reports. Deformation maps show in 2-D and 3-D where a problem has occurred in the construction process, and the coordinates of non-conforming items can be extracted, enabling speedy corrections.

Direct outputs, no processing - The dataset captured by 3-D laser scanners are so dense and rich that some information can be extracted directly out of the point-cloud. For example, the volume calculation of a stockpile can be obtained automatically with highly accurate results, since it depends only on the data resolution.

Another application is using point-processing software to compute the volume of earth removed from a project automatically on a day-to-day basis. It compares the point-cloud from Day 1 to the point-cloud from Day 2 and gives the exact amount of material removed. Because of the high density of information, level curves are generated automatically from a point-cloud and do not require additional work in other CAD packages. Distance, angle, and slope measurements are accessible directly through the point-cloud processing software and are generated into spreadsheet reports.

In addition, some vendors offer a limited - “lite” - reducedcost version of their point-cloud processing software that allows measuring and computing direct deliverable information out of the dataset. These versions enable project reviews to be completed by someone who does not have all the skills required to process a point-cloud.

Summary Point-cloud processing software is a valuable tool as a standalone application. Direct results from the rich dataset of a 3-D laser scan can be extracted and provided directly to contractors in the field, without having to process the data. When choosing a solution, keep in mind that point-cloud processing software should be able to process data rapidly and automatically so that standard CAD packages can be used to create 2-D drawings.

Additionally, the versatility of point-cloud processing software is a key driver, as it will enable increased productivity and more efficient project completion. 3-D laser scanners are fast, and the point-cloud processing software used in conjunction with it should be fast as well to automate processes and create deliverables quickly.

Omar-Pierre Soubra is a portfolio manager, 3-D Scanning, for Trimble. Trimble is a pioneer in 3-D laser scanning for civil engineering applications, having introduced the first system in 1992 through its MENSI subsidiary. Soubra can be reached at omar_soubra@ trimble.com.