- About Us
20 January 2010
US-based specialist construction writer Richard Ries looks at the science and technology behind Topcon’s latest survey mapping solutions. Note that this was written for a general consumer audience, rather than specialised industry people.
With the map features found on many Internet sites today, you can zoom into your country cousins’ property and see if they planted corn or soybeans this year. If it’s corn, you can almost see the silk drooping from the ends of the ears. If it’s beans, you can almost see the fuzz on the pods.
How on Earth can these sites provide this level of visual information about any place on Earth to you wherever you are on Earth? Is it magic?
Well of course, it isn’t magic, it’s technology – although the line separating the two almost grows less distinct every day.
Truth is, the base technology in this case isn’t all that magical.
According to Charles Rihner, vice president of planning for emerging business units at Topcon, it is a collection of familiar, proven components that form the foundation of the system.
Those components are satellite navigation systems, cameras, lasers, a vehicle’s onboard computer (CANbus), and a laptop computer.
But tying together the data from all these sources is as close to magic as you can get.
The unifying force comes in the form of Topcon Positioning Systems’ IPS-2 parallel program measurement integrated positioning system (see the separate section at the end of this article).
“Before data can be integrated, individual data points must be gathered,” said Rihner. “This is done in real time by driving a research vehicle along the desired route.”
As the vehicle proceeds, the IPS-2’s GNSS antenna gathers position data from the two satellite constellations currently in operation, GPS and GLONASS (again, please see separate section below).
Six cameras, which are ganged together, provide 360-degree digital imaging at 30 frames per second. Each frame is time-stamped and geo-referenced.
A light detection and ranging (LiDAR) system integrated in the IPS-2 delivers 3D information about structures, tree cover, and other roadside features.
“LiDAR is like radar,” Rihner said, “but it uses lasers instead of radio waves and therefore employs shorter wavelengths.
“LiDAR also has excellent beam density and coherence. These traits provide backscatter, a highly detailed reflected image that, unlike radar, picks up small items and non-metallic objects.
“Radar is great for finding airplanes and ships, but LiDAR finds such features as leaves, fiber optic lines, and even smoke and dust,” he said.
A typical IP-S2 setup would use three laser scanners, one on the left side of the vehicle, one on the right, and one more aimed straight up or to the rear. The system will support as many as six lasers.
Running at 75 Hz, each scanner delivers 45,000 data points per second. As with the digital images, each data point is time-stamped and geo-referenced. Scanners from SICK AG of Germany are well suited to the IP-S2 application.
Information on pitch, roll, and yaw is delivered in the form of an inertial measurement unit, or IMU.
Pitch describes whether the vehicle’s nose is pointed up or down or remaining level.
Roll describes whether the vehicle is level from side-to-side.
While both pitch and roll relate to horizontal planes, yaw is the amount of rotation around a vertical axis, as when the vehicle is turning.
The vehicle’s CANbus gathers odometer and tracking information from wheel sensors, which are part of a vehicle’s anti-lock brake system.
The IMU extrapolates velocity information, which includes both speed and direction, from the data provided by these sensors. Vehicles lacking ABS can use aftermarket, add-on wheel sensors.
The IPS-2 and the laptop computer work together to co-ordinate all the information gathered as the vehicle moves about.
Information is available in real-time in the field.
Further refinement of the information takes place during post-processing, when the resulting data-rich file is combined with an analog map to create a digital map with extremely high accuracy and which includes 3D elevation views and time stamps.
The last step is to remove “noise”, such as out-of-range data points, and put the final polish on the file. Post-processing is especially adept at restoring the accuracy of data gathered using the IMU only, which occurs when satellite signals are unavailable.
Who needs it?
Okay, it’s cool to see corn tassels shimmering in the country sunshine on your computer screen, but it’s unlikely anyone would commit the necessary resources for such precise mapping just to amuse ’net surfers.
But as Eduardo Falcon, senior vice president for emerging business units at Topcon, points out, there is a broad assortment of mapping applications that justifies the investment. They include:
• Cartographers, who make up the biggest market for IP-S2, whether they’re creating maps for print publications, online customers, or on-board vehicle navigation systems.
• Surveyors, who rely on IP-S2 because the information it provides is complete, accurate, and up to date.
• Road maintenance crews, who use IP-S2 to locate and identify areas where work is needed. It not only gives the location but also pictures of problems such as blocked drains, potholes, rutting, and edge crumbling. Snow removal operators can’t see the roadway under a foot of snow and use IP-S2 for edge detection.
• Telephone, electric, and cable providers, who can see where trees need to be trimmed back from transmission lines. IP-S2 will be a key enabling technology for intelligent transport systems and the automated vehicles that run on them.
While every aspect of IP-S2 is impressive, three features truly stand out.
1. IP-S2 can use IMU data to interpolate position information when satellite signals are blocked. Signal blockage can occur when the research vehicle is in a canyon, either natural or man-made (think high-rise), or under a dense canopy of trees. This makes IP-S2 the tool of choice when gathering information in urban settings. Using an IMU, IP-S2 can continue to collect data even in tunnels and under bridges.
2. IP-S2 is scalable. Not every application needs every feature. Cameras, lasers, and other components can be added or removed as needed for the task at hand. Because the whole raison d’être of IP-S2 is to integrate data from parallel sources, custom configuration of the system really is a plug-and-play operation. As new components are developed in the future, these, too, can be added to the IP-S2 repertoire.
3. While the point cloud acquired by LiDAR is adequate for identifying objects and features, the integration of the visual images provided by the cameras adds even more value. A tree is identifiable in the point cloud, but from the camera image a technician can tell if the tree is healthy or diseased and whether it’s an elm or an oak. By clicking on that tree’s image in the point cloud the technician can access all the geospatial information stored by IP-S2.
This level of detail makes IP-S2 the preferred tool for mapping and for asset management because it saves so much time and money, said Falcon.
“With traditional survey mapping, it’s hard to gather the right amount of information.
“The surveyor may spend excessive time in the field gathering too much information for fear of overlooking something. Or the surveyor may indeed overlook something and then have to make another trip to the field to gather the missing data.
“With IP-S2, neither situation will occur. A single pass gathers all the information the technician needs, and that information is then available at the technician’s computer,” he said.
Although IP-S2 is a new development, mobile mapping began in the late 1980s.
Early challenges included limitations in both data storage and scanning technology. Those challenges have long been met, but new ones have taken their place.
The biggest problem, said Falcon, is that “the information is dumb. It still takes a human to extract and interpret and edit the data.” The next step, then, is to automate the interpretation.
It would be fairly easy to program IP-S2 to recognise stop signs, for example, except that computers don’t recognise things in the same manner that humans do.
When approaching a red, octagonal sign with the letters “STOP” on it, both the computer and the human will recognise the object as a stop sign.
But what about when the approach is from an angle?
What if the stop sign has baffles to make its message unidirectional, as is done when one road intersects another at an acute angle?
What if the stop sign is part of an ad on the side of a city bus (“STOP wasting time and money. Let ACME Corporation help with your investments”) and what if the bus is moving? The human will assess each of these situations with unfailing accuracy.
The computer’s accuracy will likely be much lower.
“Software engineers can set parameters to help interpret the data,” said Falcon. “At the extreme it’s artificial intelligence.”
Even if the algorithm can’t identify every feature with 100% accuracy, the autoprocessing of just the easily identified objects would greatly reduce the need for data interpretation by humans.
It’s no exaggeration to say that IP-S2 will revolutionise survey mapping and asset management.
IP-S2 gathers information quickly and accurately and it doesn’t miss a thing. It reveals the location and condition of every single feature it encounters.
From the tassels in your country cousin’s cornfield to the parking meters outside your city cousin’s office complex, IP-S2 sees everything and stores everything it sees.
It isn’t magic, but it’s close. Really close.
Footnote 1: IPS-2 is the second generation of a parallel program measurement system used to monitor multiple programs that are running simultaneously.
Essentially, IPS-2 is a “traffic cop”. It makes sure the data tributaries from the various components of a research vehicle flow smoothly into one wide data stream.
It allows programmers to easily identify data conflicts and bottlenecks so they can be resolved.
IPS is built on the Charlotte Distributed Operating System. This second iteration extends the original IPS system with new instrumentation techniques, an interactive and graphical user interface, and new automatic guidance analysis techniques.
Footnote 2: GNSS (Global Navigation Satellite System) is the preferred term when discussing satellite positioning systems.
GPS, the 24-satellite constellation deployed by the US government, is part of the GNSS, as is Russia’s GLONASS system of 17 satellites. The European Union’s Galileo system is to be operational by 2013 and China plans to expand its regional Beidou system to a global system, named COMPASS, by 2015. GNSS is the collective term that includes all of these systems.
Proud to be partnered with: