• Air Toxics
• CMAQ
• Conformity
• It All Adds Up

• Laws and Regulations
• Maps
• Policy and Guidance
• Research
  • General Practice
  • Air Quality Planning and SIPs
  • Emissions Analysis
  • Land Use
  • Multi-Jurisdictional (Complex) Areas
  • Project Level Conformity
  • Small Urban and Rural Areas
  • Tools & Methodologies
  • Transportation Control Measures
• Resources
• Models & Methodologies
• Training
• Air Quality Newsletter

Improving Vehicle Fleet, Activity, and Emissions Data for On-Road Mobile Sources Emissions Inventories

Appendix B: Guidance on Conducting License Plate Survey

Introduction

In the development of on-road mobile source emission inventories, reliable data on vehicle fleet characteristics such as age distribution and alternative vehicle fuels and technologies fraction are as important as accurate data concerning vehicle activity and emission rates. For example, fleets with a higher percentage of older vehicles will have higher emissions for two reasons. First, older vehicles have typically been driven more miles and have experienced more deterioration in emission control systems. Second, a higher percentage of older vehicles also means that there are more vehicles in the fleet that do not meet newer, more stringent emissions standards. Studies have shown that the class and age distributions of vehicle fleets can vary significantly from area to area[1]. Therefore, for state implementation plan (SIP) and transportation conformity purposes, the U.S. Environmental Protection Agency (EPA) recommends that agencies develop local vehicle class and age distributions[2].

The current practice in developing vehicle fleet data relies heavily on the use of vehicle registration database. This practice is sufficient for creating base vehicle fleet characteristics for an area, but it also has certain limitations. For instance, a significant fraction of vehicle miles traveled (VMT), and thus emissions, for an area could be attributable to vehicles registered outside the area[3]. This is usually the case for, for instance, tourist areas and areas with major transportation hubs such as international airports and seaports. Applying inaccurate vehicle fleet characteristics could have significant impacts on the resulting emission inventories.

Furthermore, some vehicle emission and air quality studies may benefit from careful characterization of vehicle fleet at a more refined scale. For example, a project-level analysis of vehicle emissions may be performed using vehicle fleet data localized to the project. In another example, a modeling of regional air quality during ozone episodes may be based on emissions estimated using vehicle fleet data specific to the modeling period. Yet, in another example, a measurement of pollutant concentration near roadways for dispersion model development or validation may be conducted in conjunction with vehicle fleet characterization during the measurement in order to reduce biases in the emission estimates.

The need for better vehicle fleet data mentioned above can be met by conducting vehicle license plate survey. The survey is aimed at collecting the license plate numbers of vehicles that are present at the location of interest during the survey period. There are two types of license plate survey that may be conducted--on-road survey and parking lot survey. This guidance is focused on the on-road survey. It provides guidelines for selecting survey sites, videotaping vehicular traffic in the field, and extracting license plate numbers from videos.

Selecting Survey Sites

There are two aspects to the selection of license plate survey sites the general area where the survey will be conducted and the location of the camera set up and videotaping. These two aspects should be considered together. The considerations include:

The general area where the survey will be conducted should be selected according to the purpose of the study. For example, for a project-level emission analysis the survey sites should be at or within the vicinity of the project. For a regional emission inventorying purpose, the survey sites should be on major roadways in the region. Figure B-1 shows an example of survey sites that are appropriate for a study to examine the characteristics of out-of-region vehicles. In this example, areas on major highways that carry traffic into and out of the region around the boundary of the region are selected.

Figure B-1. Survey sites for out-of-region vehicles study

• If there are multiple roadways in the general area that has been selected for conducting license plate survey, the roadways that carry high volume of traffic are preferred for two reasons. First, having high volume of traffic means more number of license plates can be captured within the same amount of time. Second, vehicles on a roadway with high volume of traffic are more likely to travel at a slower speed, which makes it easier to read the license plate numbers.
• The location of the camera set up and videotaping can be on roadside or on a bridge overcrossing the roadway. If available and suitable, the latter is preferred for two reasons. First, it is safer. Second, unlike a roadside location from which the view of the traffic in the inner lanes may be occluded, a location on a bridge allows for the traffic in any of the lanes to be captured.
• In the case of a bridge, the surveyor should find one with a wide shoulder lane or sidewalk that provides enough working space for setting up video cameras and other equipment (see the left photo in Figure B-2). Also, most bridges have fences on either or both sides of the bridge. These fences may be made of wire mesh, steel bar, or similar material. A bridge without fences or with ones that have large opening spaces is preferred (see the right photo in Figure B-2).

Figure 2. Location of camera set up and videotaping

• In the case of roadside, the surveyor should find a section of the roadway with a sidewalk, an elevated bank, a wide median, or a safe spot away from the traffic.
• Depending on the survey location, a permit may be required in order to occupy the site and videotape the traffic. The permit may be acquired from the responsible agency such as the City (for city streets), the County (for county roads), and the state department of transportation (for highways). The permit application and approval process may take days or weeks. Therefore, the surveyor should apply for it as soon as the survey location has been identified.

Videotaping Vehicular Traffic

Equipment

The equipment required to perform a successful on-road license plate survey can vary by survey site, but the necessary items are listed below:

• Video Cameras - Choice of cameras is especially important when videotaping vehicular traffic. A high definition (HD) camera gives better resolution of the images than a standard definition (SD) one. A strong optical zoom (as opposed to digital zoom) is needed to get a clear view of the license plates. Storage space is another consideration, as cameras may have to hold multiple hours of video at any point in time.
• 12V Marine Battery - When using the videotaping mode with the highest quality (typically the standard play or SP mode), stock camera batteries provide approximately 45-60 minutes of filming time. If the survey session(s) for the day is longer than that, additional power will be needed. A marine battery can provide sufficient, stable power to run two HD cameras in SP mode for at least eight hours.
• 12V to 120V DC-AC Converter - This adapter is required to connect the cameras to the marine battery.It is still recommended to leave the stock batteries in the camera in the event that this adapter fails.
• Tripods - A standard tripod is acceptable, but tripods with thicker, sturdier legs are preferred. Smaller tripods are at risk of being blown over, necessitating careful observation or heavy anchors.
• Safety Vests - Reflective safety vests help to ensure the safety of the surveyors on site.
• Permits - An on-road survey requires the appropriate encroachment or filming permit approved by the city, county, or state.

In addition to the necessary items above, the following optional items may also be useful:

• Digital Multimeter - This is useful for keeping track of the 12V charge level.
• 12V Batter Charger - This is needed to recharge the 12V battery after the filming.
• Extension Cords - They allow for optimal positioning of the cameras.
• Traffic Cones - They are helpful for guarding the working space.
• Personal Supplies - The surveyor may need to perform the filming for several hours at a time. This means that matters such as hydration, food, and shade should be prepared for before the start of the survey.

Procedures

The procedures for videotaping vehicular traffic vary by circumstances. Below are some basic procedures and considerations that may be used as a general guideline.

• Survey Period - The videotaping should be performed during peak hours for the same reasons as performing it on roadways with high traffic volume. First, the number of vehicles is high so more license plates can be captured. Second, the traffic is more likely to be congested, allowing the license plate numbers to be clearly seen on the video playback.
• Initial Setup - At least 15 minutes for setting up and taking down the equipment should be allotted on both ends of the planned filming interval. Place the 12 V marine battery and the connecting DC-AC adapter in the center of the working space. Use extension cords, if necessary, to help provide flexibility of camera placement. Also, tie down each tripod to a centered weight between their legs to provide stability. However, tying cameras to a fence is not a good idea, especially in the wind, as it will cause video shaking and stuttering. If possible, the initial setup should be tested prior to the planned filming day to ensure that all the equipment is working properly.
• Camera Angle and Zoom Level - One of the most critical aspects to capturing good license plate footages is the shooting angle of the video camera.On an overpass, the camera should be as close to the ground as possible and zoomed to a point on one lane of traffic so that the camera view is between 30 and 45 degree angle with the road surface. Conversely, while filming on roadside, the tripods should be set at their maximum heights to properly record the plates.

Another critical aspect is the camera's zoom level. Adjust the camera such that the lane markers of one lane are on the edges of the screen (use maximum zoom if necessary). The license plate must be visible in the camera view for at least two full seconds to be properly processed. It is recommended that a trial-and-error approach be used to find an optimal shooting angle and zoom level.

During a survey session, cameras should be checked every 15-30 minutes for shooting angle, zoom level, as well as power. As the lens is set to near its maximum zoom, any slight adjustment can take the focus completely off of the license plates. In some cases, the surveyor may contact the camera or the tripod accidentally, which can make the shooting angle and zoom level no longer optimal.

• Camera Glare and Heat - Morning and late afternoon glare can be very problematic to an on-road license plate survey. The most effective solution is manual readjustment of the "exposure" setting on the camera. This could almost entirely eliminate glare, but needs to be checked every 15-20 minutes as the position of the sun changes.
• In addition, video cameras can be prone to overheating if being used in the sun unprotected. This is generally not a problem, but in particularly hot areas, cameras can get too hot and temporarily shut down.If this happens, try to provide shade protection for the camera (draping a towel above or over the camera suffices).
• Camera Rotation - Depending on the number of lanes being surveyed, length of the survey, and the number of available cameras, a camera rotation scheme may be required. An example is given below.

After each survey session, there are a few routines that should be performed to prepare for the next day of survey. It is important that these routines are performed immediately after each session, especially if multiple sessions are planned for the same day.

• Battery Charging - Each camera should be fully charged as the stock battery serves to protect against any accidental shutdowns of the main power source. Each 12V marine battery should also be charged overnight to provide enough power for the next day of filming (charging mid-day is not necessary).
• Unloading Data - Even cameras with high capacity data storage have a fairly limited recording time. Every time the equipment is packed up, the data stored on the camera should be unloaded onto an available computer or external hard drive.

Extracting License Plate Numbers

Due to the high costs of automated license plate recognition systems, manual extraction of license plate numbers can prove to be a viable alternative. A typical setup involves a team of data technicians; each assigned a set of videos from the survey. The process is straightforward; going through the tape, pausing at each vehicle, and recording the vehicle information on a spreadsheet. If possible, a dual-computer or dual-monitor setup such as that shown in Figure B-3 is desirable as it will help make the processing faster.

Figure B-3. Dual-computer setup for license plate number extraction

The main advantage of manual extraction is the interpretation ability of the data technician. Due to the relatively ad hoc nature of the data collection process, the size of a license plate number, the amount of lighting on the plate, and the amount of time the plate is captured on film can vary greatly from one video to another. The human eye, however, is quick to adjust and able to interpret license plates. Moreover, the manual extraction technique allows for the collection of more vehicle attributes than just the license plate number, for example, state of registration, vehicle class, or even vehicle make and model in some cases.

On the other hand, manual extraction takes time. Even with a dual-monitor setup, the amount of time to process one hour of video can take between two and four hours depending on the traffic volume and the quality of the video images.

[1] Malcolm, C., Younglove, T., Barth, M., and Davis, N. (2003). Mobile-source emissions: analysis of spatial variability in vehicle activity patterns and vehicle fleet distributions. Transportation Research Record, 1842, 91-98.

[2] U.S. Environmental Protection Agency (2009). Technical Guidance on the Use of MOVES2010 for Emission Inventory Preparation in State Implementation Plans and Transportation Conformity. Report No. EPA-420-B-09-042, December.

[3] Lutsey, N. (2009). Assessment of out-of-state truck activity in California. Transport Policy, 16(1), 12-18.