The University of Michigan's football program is arguably one of the best in the nation. With a 126-year history, the U-M Wolverines hold the record for the all-time Division I-A leader in wins (842) and for the nation's longest active streak of bowl game appearances. With such an impressive program, the team deserves an equally impressive stadium. The "Big House," as it's called, is the nation's largest collegiate football stadium with a capacity of 107,501 (at least 110,000 fans show up to each game). The stadium has led the nation in attendance for 28 of the last 29 years, and on Saturdays during football season, the stadium becomes the fourth largest city in the state. This impressive structure has undergone many renovations in its 78-year history while maintaining its record as the largest stadium. With many more renovations planned for the future, university officials are in need of accurate and up-to-date as-built plans of the existing structure. This was the job Midwestern Consulting had been hoping for.

This true color point cloud contains approximately 130 million points and shows the entire area Midwestern Consulting scanned.

Presenting the Technology

Midwestern Consulting is a civil engineering and land surveying company with more than 70 employees located in Ann Arbor, Mich., just a few miles from the stadium. We have always been committed to implementing new technology in our operations. In 2002, we purchased our first laser scanner, a Cyrax 2500 (Leica Geosystems HDS, San Ramon, Calif.). Last fall we upgraded our scanner to a Leica HDS 3000 with a 360° horizontal x 270° vertical field of view. We dreamed of a project that would best exemplify the cost savings and efficiency made possible by the panoramic field of view, and knew the University of Michigan football stadium would be that project. Too bad for us that the university didn't need any survey work done "¦ or so we thought.

"When the request for proposals came out [from the university], we could hardly believe it," says Patrick Hastings, PS, principal in charge of survey operations at Midwestern. "Nobody expected the university to undertake stadium renovations for several years." The university needed a site topographic survey and additional building surveys prepared inside and around the stadium. "We sat down and looked at the benefits of using traditional surveying methods versus using our laser scanner, and calculated the cost of each," Hastings says. "We estimated the cost between the two methods to be nearly identical. What made us decide to use the laser scanner, among other things, was that the requirements for the survey were to locate all accessible structures. With the laser scanner, everything was accessible." Midwestern was, however, concerned that our price might be higher than the competition so we decided to invite the engineer in charge of stadium renovations to our office for a demonstration of the scanner's benefits. A few days later, we received notice that two engineers would be coming to our office. Rather than showcasing one of our previous projects, we quickly headed to the stadium and did a sample scan. We then created a demonstration for the engineers. They were impressed and agreed that there were some real benefits to the laser scanning technique, such as documenting all structures safely from the ground. However, even though they were impressed, they were not willing to award the entire project to us just yet. Instead, they asked us to do a pilot project on half of the seating bowl.

Midwestern provided U-M with this true color point cloud of the south end zone scoreboard. The client needed elevations of the video matrix of the scoreboard and the top of the scoreboard.

Scanning the Big House

Our first task was to survey all of the walkways in 23 of the stadium's 44 sections. The university plans to replace the concrete in the lower seating bowl this year and it wants to be sure that the steps and seats are reconstructed to their original location. Our plan was to set the scanner in the press box for a panoramic view looking down on the stadium.

We set up the scanner, and while it went through its initialization process, we set up the control points (scan targets). These targets were located using a Leica Geosystems TCRA1103+ total station (Leica Geosys-tems, Norcross, Ga.), which not only gives real-world coordinates to the scan data but also serves as an independent check on the scanner's accuracy. Once everything was set up, we scanned a target to serve as a backsight check. When the scan was complete, this target was scanned again to ensure there were no errors with the scanner or its setup while it was scanning. This procedure wasn't always needed with the Cyrax 2500 because most of those scans lasted less than 30 minutes. The increased field of view of the HDS 3000 means that scans could last much longer and could be more prone to errors. Our first scan from the press box lasted almost four hours and contained 20 million points, but it had some errors. When the backsight target was scanned at the end of the scan and compared to the initial scan, it had moved 0.2 feet, much more than the 0.02 feet expected from previous experience. After we checked to make sure the target hadn't moved and rechecked the scanner setup to ensure nothing had slipped and it was still level, we were a little frustrated. We consulted Leica's HDS support, and determined that slight movements, due to factors such as thermal expansion or the tripod settling (the scanner weighs 35 lbs), can have an effect on the scan data. Unlike modern total stations, the HDS 3000 doesn't yet have a vertical or horizontal axis compensator, so even slight movements can greatly affect the results. To remedy this, we limit scan times, which in turn limits the slight movements that may occur. A setup may still take four hours, but it will be comprised of many individual scans. Still, it only took two days for one scanner operator to scan the 23 walkways from the press box and from within the stands. In all, the scans generated approximately 30 million points. But the university did not want just a point cloud; they wanted traditional drawings showing cross sections through the stadium.

The two days of scanning created 19 point clouds that were aligned (registered) to each other and to the site coordinate system using Leica Geosystems HDS Cyclone software. Once we had a complete point cloud, we imported the data into AutoCAD (Autodesk, San Rafael, Calif.) using Leica Geosystems HDS CloudWorx software for AutoCAD. There, the point cloud could be quickly sliced to isolate each walkway, and cross sections of the steps were drawn. For our pilot project, we met the engineer's request of creating drawings showing cross sections with the elevation of each fifth step and horizontal dimensions of the treads. After the university received and reviewed our drawings, they changed their criteria and wanted elevations on specific steps, not every fifth. If we had used traditional field methods, we would have shot just what they had asked for, every fifth step, and would have had to send a field crew back to the site to re-shoot the specific steps they later requested. However, since we had a computer representation of the stadium, we simply had to revisit our files to get the data they needed.

The panoramic field of view of the HDS3000 allowed this point cloud to be created from one setup.

A Chilly Obstacle

We began surveying at the end of the football season in November, but the snow so familiar to Michigan soon interrupted our work. This past winter was one of the snowiest in recent history, and our ability to use the scanner was debilitated since the unit cannot be used when there is precipitation. Not only is the scanner not water-resistant, but it picks up rain, snow and fog in the air, causing noise in the data. When there was not snow on the ground or in the air, our one-man scanning "crew" was onsite at the stadium. Often, while the scanner was running, he could accomplish other tasks such as operating the total station to locate targets or making hand measurements to verify the scan data. When all peripheral work was done, there was time to sit in the truck and warm up.

"Being able to operate the scanner via a wireless connection or the 100-foot cable was a real advantage," says Charles Huber, laser scanner technician. "I could establish the area to scan, set up my scripts and monitor the progress of the scanner without standing next to it." Scripting was a big time-saver since we could specify different point densities over the scanner's field of view without having to spend time scanning the entire field of view at a high density. Once the script was created, the scanner could perform multiple operations without user input.

Ready for Renovation

Overall we took approximately 600 scans from 85 locations around the stadium. All of these scans were registered into one complete point cloud that was used inside AutoCAD to create the 2D cross section and floor plan drawings the university requested. On many occasions, the university needed information in addition to what was shown on our drawings, such as dimensions between the scoreboards, the location of the end zone lines and additional cross sections through the adjacent basketball arena. Their requests were usually answered the same day since we already had accurate data that gave us all of this information.

We were pleased that our pilot project evolved into scanning the entire structure, the whole seating bowl, the outside of the stadium and the enclosed underside. Surveying the entire structure will facilitate renovations that may include a new press box, enclosed skybox seating, a new concourse, changes to all seat widths and restroom upgrades.

While these renovations are transforming the "Big House" into the "Really Big House," the engineers and architects can be assured that the as-built drawings are complete and accurate thanks to the latest in 3D laser scanning technology. Additionally, the power of a GIS will allow planners to maximize the number of seats in the stadium to make sure Michigan fans can still take pride in being the largest college-owned stadium-a grand home for a leading football contender.

Sidebar: Benchmarking a GIS

There was one additional request the scanning data did not cover: the length of every bench in the stadium. The university wanted this information to help them reconfigure the facility's seating and create a database of all 107,501 seats. We accomplished this request using two robotic total stations, a Leica TCRA 1103+ and a Trimble 5600 (Trimble, Sunnyvale, Calif.), each set up at one end of the stadium. Over three days, two surveyors walked the entire snow- covered seating bowl and shot 8,000 points on the ends of the benches all accurate within a 1/4". The university supplied us with its current listing of seat numbers per bench and we turned to our GIS expert to put it all together. " I don't know of a project like this ever happening before," says Ben Chen, PE, senior project manager. "The amount of information was overwhelming, yet perfect for a GIS." Once the bench lengths and number of seats were input into ArcGIS (ESRI, Redlands, Calif.) we were able to create thematic maps showing the seat widths per bench and the overall average seat widths. " Now whenever the university wants to try a different seating configuration, they call me and within a few hours I can tell them what effects it will have on the seat count and seat width."