A 902Kb PDF of this article as it appeared in the magazine complete with images is available by clicking HERE
Terrestrial Mobile LiDAR Scanning, otherwise referred to as mobile mapping, is an effective tool helping DOTs, engineers and surveyors contend with our aging transportation infrastructure. Mobile mapping platforms collect millions of range measurements on roadways and surrounding surfaces while driving at highway speeds. As a result, detailed 3D surface geometry data is collected at unprecedented rates and quality. The challenge that users of mobile mapping technology now face, however, is the timely compilation of data into information deliverables.
The heart of this problem resides in scale. The benefits of mobile mapping technology are the most compelling when the scale and complexity of the project are great. Raw data, the billions of range measurements that these systems collect, is all but unusable to the end-consumer who is looking to understand something specific about their highway. In addition, the strengths of multipurpose 3D software packages, namely a broad set of features for data manipulation, make them less suitable for handling large data sets and repetitive tasks.
The team at Allpoint Systems, a 3D software company in Pittsburgh, Pennsylvania, saw this problem as an opportunity to apply their robotics knowledge to boost the productivity of mobile LiDAR technology and harness the data’s true potential. "We’ve been tackling automated processing of 3D data for nearly a decade to build robots that could interpret the environment and make informed decisions" says Seth Koterba, VP of Technology at Allpoint. "Many of the same tools and concepts that enable a mobile robot to navigate autonomously can also address a modern surveyor’s needs."
This theory was successfully put into practice when, in the fall of 2010, the survey firm Terrametrix (a Terrestrial Mobile LiDAR Scanning service provider based in Omaha, Nebraska) began a bridge clearance pilot with the California Department of Transportation (Caltrans). The Terrametrix mobile mapping system was capable of capturing 100 to 400 bridges in a week depending on the density of bridges, traffic, and weather conditions. Their ability to process data into reports did not, however, match this collection capacity. "When we discussed the opportunity with Terrametrix, we discovered that their daily deliverable production capacity was around two to four bridge clearance reports per person," stated Koterba as he recounted their initial conversation. "We were confident that we could do better."
Three months, 589 bridges, 1,000 miles, 8,000 measurements, 1.5 billion points and 1 terabyte of data later, the final report was delivered and the performance findings made a compelling case for the utilization of automated 3D processing. Production rates for bridge clearance reports by a single user went from 4 reports a day to 100 reports per day–a 25X improvement! For every day in the field spent collecting, reporting time went from 12 days (manual) to just 1 day (automated).
The technology empowering this automated software resides in robotic "perception" algorithms, which provide rapid classification, segmentation and verification of extracted 3D information. Features of interest, such as bridges and road surfaces, can be specified and extracted from large 3D data sets without direct user interaction. As such, significant efficiency gains are possible by enabling user workflows that reduce non-value operations. According to Ryan Frenz, lead developer of the perception system at Allpoint, "So much time was being put into repetitive low value tasks, such as rotating and translating 3D data on the computer screen. Pay attention to how much distance your hand has move or how many mouse clicks it takes to extract a cross section from a 3D point set. Eliminating these repetitive low value tasks streamlines effort so that the user can focus on higher value activities such as quality control."
The culminating outcome of this union between surveyor, mobile LiDAR and perception software is the rise of small survey teams capable of completing extraordinarily large projects quickly. Today, for example, Terrametrix is tackling California’s state inventory of bridges and overpasses.
Efficiency gains are not the only advantage. Automation also means achieving a level of consistency and repeatability that’s almost impossible for a human counterpart–bolstering the other benefits of LiDAR based solutions. Because every step in the process is digitally executed, it’s also documented– providing a clear chain of custody. This also offers the ability to re-process entire datasets programmatically if requirements are altered; as may be the case if new regulations are introduced requiring different or additional information to be extracted from the datasets.
Bridges are only the beginning
LiDAR technology is evolving rapidly, lowering the speed and cost of collection; more data is being collected than ever before. Automated processing means that we can get higher throughput and even obtain additional results from the same data, still with time to spare.
Do more with data that’s already at our disposal.
We’re now in a position where it’s more efficient to collect a large amount of data and use perhaps only a small portion of it to make a decision–the equivalent of taking a video when you only need to deliver a photograph. What other valuable information can be extracted from that data?
Take on new projects that may have been uneconomical just a few years ago.
Having access to more powerful hardware and automation software decreases the scope and/or risks for surveying projects. This means that surveyors can afford to be more flexible in the size and types of projects they accept.
Provide previously unimaginable solutions
We’re just starting to scratch the surface of what’s possible with automation. Time marches on, and as we enter the efficiency phase of one technology we’re also collectively enabling completely new adoption curves, markets (or sub-categories), and adopters. Just as LiDAR opened the door for mobile mapping, and both presented an opportunity for automation, we continue to progress and discover new capabilities. The data’s there, the capability is there, what else can be done? How else can automation be used to improve workflow? What other information, such as feature identification, additional clearances, object properties (surface area, volume) can be extracted from datasets?
Allpoint Systems develops automated software inspired by robotic technologies that, in one of the first applications, yielded a 25x efficiency improvement in reporting throughput. At Allpoint and Terrametrix, we’re constantly asking and enthusiastically answering questions like these as we help clients do more with their time.
Aaron Morris holds a Ph.D. in robotics from Carnegie Mellon University and is the CEO and founder of Allpoint Systems, LLC. Allpoint’s software solution, the Perception EngineTM, was developed to make possible data deliverables that would be difficult to do through manual processing. Their experience is in creating a hybrid software toolsets that mix automation (for extremely tedious tasks) with streamlined user interaction (for efficient high-level decision making).
A 902Kb PDF of this article as it appeared in the magazine complete with images is available by clicking HERE