Video-based roadway asset inventories: technology streamlines a labor-intensive task
Gordon Daring, P.E.
Infrastructure Asset Management Team Leader
Vanasse Hangen Brustlin, Inc.
Chair, APWA Engineering and Technology Committee
Most public works managers today recognize the value of having an accurate inventory of the infrastructure assets they've been entrusted to maintain—it's difficult to manage and maintain something if you don't know where it is or what condition it's in! Further, most public works departments recognize a geographic information system (GIS) as the ideal platform in which to store such inventory data. Still, today there are a great many public works departments that do not have an accurate GIS-based inventory of their infrastructure assets.
Historically, the effort involved with developing an accurate GIS-based infrastructure inventory has been labor and time intensive, and therefore expensive. This situation has improved with the growing availability of systems designed to extract roadway feature inventories from digital videos.
In the 1970s, transportation agencies began developing roadway video logs, enabled by a new technology at the time—the video cassette tape. The video cassette provided a relatively convenient media upon which to store and retrieve hours of video. The video logs included a text overlay to indicate the location of the video frame along a route, generally tied to a milepost or other linear referencing system. The video logs provided the benefit of being able to "drive" any road from the convenience of the office. As technology advances have been made, these systems have become more capable, convenient and affordable. The applications of today's systems have expanded to include:
Today's systems incorporate video, position, measurement, and data base functions. Typical system components include one or more high-resolution digital cameras, global positioning systems (GPS), electronic distance measuring instruments (DMI), inertial navigation systems (INS), specialized data capture software, video playback and analysis software, and GIS. Some systems include pattern recognition software, capable of recognizing specific features such as stop signs.
Most often, one or more cameras are mounted on top of a high-profile vehicle (van or SUV). Cameras may also be mounted inside the vehicle, and aimed out through high windows. The cameras may be focused in any direction, depending on the intended use of the video data. If the video will be used to extract roadway feature inventory data, the cameras will be focused forward of the vehicle. Side-focused cameras can be used to capture images of buildings, property, and other features away from the roadway. The forward-focused cameras can be used to evaluate pavement condition. In addition, some systems include the capability to utilize downward-focused cameras that provide a clearer view of pavement distresses.
The cameras used are high-resolution progressive scan digital cameras. Still images are captured from the video and stored as .avi files in real time. The still images are captured at a user specified interval, generally in the range of every 10-50 feet. When the images are played back, they give the impression of a continuous video. The storage of the video data as a series of still images is critical to the ability to use the data to accurately locate roadway features.
The coordinate location of the vehicle is always known, either from GPS or from the inertial navigation system, which continues to track the vehicle location if the GPS signal is lost. The location and angle of each camera is also known through a careful calibration process performed prior to conducting the video survey. With most systems, as long as a feature can be seen in two video images, the feature's coordinate location can be determined. The location of a feature is determined by an operator in the office viewing the two images through special video analysis software. The operator clicks the location of the feature in both video images, and the software computes the feature location using triangulation. This approach typically results in sub-meter coordinate accuracy for each feature. The video analysis software can also be used to measure features visible in the images.
There are a number of firms that specialize in providing video data collection services and software, as well as engineering firms who provide these services. The same firms are available to process the videos and provide roadway feature inventory data in GIS ready format. It is also possible to have the video data collected, then purchase the video analysis software and process the data with in-house staff.
Gordon Daring, P.E., can be reached at (800) 927-4363 or firstname.lastname@example.org.