It’s happened quickly. People in professional disciplines and the general public have come to experience — and expect — increasingly sophisticated tools and applications for geospatial information. LiDAR and digital imaging are major drivers behind this change. In just a few years, technologies for scanning and photography have matured to become accepted and reliable sources for geospatial data. Now a routine component of the data collection process in many industries, LiDAR and imaging are producing enormous volumes of 3D georeferenced data.
As LiDAR technology continues to evolve, the focus has shifted from gathering data to efficiently managing, analyzing and utilizing the large 3D datasets collected by airborne and terrestrial scanners. Solutions providers have created well-defined workflows to visualize the point clouds and develop them into computer-aided design (CAD) models to be used farther down an enterprise’s chain of analysis, design and decision making. These processes are familiar in applications such as construction, architecture, industrial plant operations, facilities management and more. Over the last couple of years, much of the technological development in LiDAR has centered on how to use point clouds to automatically produce 3D CAD models.
Similarly, development in imaging and photogrammetry has been very dynamic over the last few years. A number of companies now provide solutions for collecting imagery using unmanned aircraft systems (UAS) or terrestrial cameras and using it to create either photogrammetric or 3D mesh deliverables. For the most part, however, imaging has yet to evolve in terms of providing data that can efficiently be converted into CAD models needed by engineers, architects and contractors. That situation may be changing.
Visualization: The First Step In Modeling
A common approach of LiDAR is to “capture everything in sight” and then assemble the data into a point cloud. Technicians use processing software to combine data from multiple scans into a cohesive data set, removing noise, superfluous points and outliers as part of the quality control process. People can then use viewers to visualize the project and extract individual points or make simple measurements between points in the data set.
A similar flow exists for photography and the work of using imagery to produce a 3D mesh. The objective is to enable the user to see and understand the scene and take some basic measurements. Early tools for visualizing point clouds and imagery produced 2D+ representations similar to the Streetview functionality in Google Earth. These approaches work well for many basic applications, but they are not suitable for modeling or more advanced needs that require visualization in 3D.
One effective way to provide 3D visualization is the use of simple applications to view the data. Rather than using the sophisticated software that technicians use to process the point cloud, applications known as “viewers” provide simple tools that enable multiple stakeholders to see the data in 3D. Software developers often provide these viewers at no charge, enabling geospatial service providers to deliver a 3D data set together with a viewer to their clients. Stakeholders can use the viewer to move around in the point cloud and see the project from optimal viewpoints.
More recently, Internet technologies have emerged to provide Web-based viewers and tools to further increase access to visual information. Because the data resides in the Cloud, Web viewers ensure that all users have the same data and tools, eliminating the need for service providers to send datasets to multiple clients while ensuring that stakeholders have up-to-date information. And by controlling access and permission levels, system managers can prevent accidental or unauthorized changes to the data.
For image data, Web-based viewers can provide orthomosaics and 3D representations developed from images captured using total stations or imaging rovers. Like point clouds, it’s possible to measure distances and points within the images. In addition to enabling people to visualize a project or site, photographs can be valuable contributors to the process of 3D modeling.
3D Modeling: Creating Concise Representations Of Data
In its early years, CAD served as a drafting tool to increase productivity in creating 2D plans on a computer. CAD has since moved into engineering design and 3D modeling to become a fundamental tool in many disciplines. Applications such as building information modeling (BIM), industrial facilities and complex architecture now routinely use the advantages of 3D modeling and visualization. We can attribute much of this trend to the technologies that gather, process and analyze point clouds and imagery. The information produced by these systems serves as the raw material needed for visualization and modeling, which in turn feeds the modern 3D CAD systems.
The initial step in modeling uses a LiDAR-produced point cloud to make more detailed computations and analysis. At this stage the point cloud begins to move into the world of 3D CAD and modeling. Technicians can capture dimensions and surfaces for use in planning, engineering and construction. One example is the Trimble Scan Explorer (TSE) plug-in for SketchUp Pro. Using TSE to work on a point cloud, users can extract points, corners, edges and planes. The entities go directly to SketchUp Pro, where they are used as guidelines and surfaces in the modeling process.
The next level of modeling adds the ability to create 3D CAD objects (known as primitives) that represent features such as pipes, tanks and structural components. This is typically done by selecting a group of points in a point cloud and then fitting a defined shape to it. Although the process can be labor intensive, the resulting 3D models provide significant value and time savings to downstream users.
To reduce modeling time, developers envision automated modeling systems that can streamline the process of recognizing a feature and creating the appropriate 3D CAD object. Common or standardized objects can be selected from a catalog and placed into the model based on points and measurements from the point cloud. Using software such as EdgeWise, piping can be modeled quickly by comparing dimensions in the point cloud to standard pipe diameters. Similarly, structural elements such as I-beams can be identified using databases that contain information on nominal components and sizing.
Intelligent modeling can expand to include GIS-type information on facilities and assets. To illustrate this, consider a valve in an industrial plant. Rather than manually modeling the valve from the point cloud, technicians can insert a predefined 3D model of the valve from a catalog. Along with information on the valve manufacturer and model, the system can contain attributes such as serial numbers and specifications. Using 3D visualization, users can see the valve from different viewpoints and click to drill down to access maintenance records and lifecycle information.
As 3D modeling systems continue to evolve, capabilities for automated modeling will increase. In some cases, however, modeling may not be needed. For geospatial professionals, the visualization and modeling deliverables they provide can vary based on specific client requirements.
Multiple Clients Present Multiple Needs
Providers of 3D data services (scanning, digital imagery, data processing and modeling) often describe three broad classes of clients for LiDAR and imaging data. One class relies mainly on visualization and basic measurements from orthoimages or point clouds. This work often engages planners, architects and engineers on civil and building projects who can utilize 2D+ views to develop and present concepts to stakeholders.
A second group requires 3D objects for use in their CAD systems. For example, many design engineers prefer a CAD model over a point cloud. Because modeling often requires specialized skills and software, these clients may look to the service providers to do the work of producing 3D models from the point cloud. This can be a challenge for service providers; modeling different types of industrial facilities calls for a range of specialized software and technical expertise.
A third group of clients looks to the service provider to gather and process the data to produce a point cloud. They then incorporate the cloud into their design work. These clients — typically in the industrial segments — must deal with highly complex scenes and projects. For them, it’s more effective to train people to work directly with a point cloud than to spend the time developing the point cloud into a CAD model.
The Value Of Visualization And Modeling
The ability to visualize, measure and model delivers immediate and tangible benefits. For example, detailed information on existing conditions is essential for designers planning a retrofit for a building or plant. This information provides two benefits in managing the costs of the retrofit. Design and construction teams can optimize the design to control capital costs of installation. The detailed 3D data also enables them to plan and manage the work to minimize plant downtime.
To plan the modification, teams need to completely understand the existing conditions, where the new design is going to fit, and what other parts of the facility will be affected by the retrofit project. A number of BIM tools enable them to visualize the project, even to the point of moving parts and equipment through the plant. By identifying trouble spots in the model before the work begins, the plant can avoid costs of delays and rework during the project.
Other industries can use 3D visualization and modeling in a similar way. For instance, photographic information can be used to identify and plan bridge maintenance programs, including checking clearances for cranes and other large machines. Scanning and photogrammetry can also provide data to visualize and model railway tunnels and stations as part of the work to increase safety and capacity of existing infrastructure.
Regardless of how geospatial data is collected, its value is not realized until it can be put to work. Information must be delivered to people in forms that are easily understood and readily useable. For geospatial professionals, visualization and modeling can be some of the most important and valuable services that they can provide.