Specifications

1
1. INTRODUCTION
Commercially available instruments for road-side data collection take highly limited
measurements, require extensive manual input, or are too expensive for widespread use.
However, inexpensive computer vision techniques for digital video analysis can be applied to
automate the monitoring of driver, vehicle, and pedestrian behaviors. These techniques can
measure safety-related variables that cannot be easily measured using existing sensors.
This project implemented and tested a digital video analysis system for use at intersections based
on a single camera design. As part of the research effort, a prototype video recording system was
developed for field deployment. It was designed around the objective of recording high-quality
video at an intersection for up to one week without any intervention. The design of this system is
described in Section 2. To simplify the analysis of the recorded video, a standard camera–
intersection configuration was adopted. This configuration is described in Section 3. The
algorithms used to track objects in the digital video and their implementation are described in
detail in Section 4. Video was recorded at both urban and rural non-signalized intersections and
was analyzed using the digital video analysis techniques developed in this project. These
intersections and analysis results are described in Section 5. A graphical computer interface was
implemented in order to simplify and automate the data analysis and the data review process. A
step-by-step guide to the usage of this software is provided in Section 6.
2. DESIGN OF A VIDEO RECORDING STATION (VRS)
The primary advantage of digital video analysis is the ability to automatically track vehicles and
quantify the frequency of events over a long period of time. In this research, a low-cost digital
camcorder is used to collect video from non-signalized intersections. The recorded video is used
to test the feasibility of using digital video analysis for traffic safety studies. The video taping
sessions were limited to a maximum length of one hour due to the storage limits of current
camcorder technology. Because the recording of video does not require any attention, it would
be desirable to have the capability to record video for much longer periods of time.
2.1 Design Constraints
In order to develop this capability, a portable video recording station was designed around the
objective of recording high-quality video for up to one week (168 hours) without any
intervention. This design constraint was given the highest priority. The second most important
design constraint was to find a low-cost solution through the use of consumer-grade off-the-shelf
parts and open-source software. While satisfying these two objectives alone is not particularly
challenging, the design problem became much more challenging when a third design constraint
was enforced. Because power is not typically available along the road or at all intersections, we
required that the station be completely self powered for the one-week recording period.
Satisfaction of these three design constraints made the problem much more difficult, as it forced
consideration of the tradeoffs between low-power and low-cost hardware as well as the tradeoff
between battery and solar power. A number of additional design constraints were enforced.