U09: License-Plate Recognition (Phase B)
In April 2006, Knoxville, Tennessee joined an increasing number of cities in reducing speed limits for large-trucks (with gross weights over 10,000 pounds) on the interstate highways in its metropolitan area. In recent years, reducing large-truck speed limits in urban areas has become one of the preferred countermeasures for combating urban air-quality problems. The rationale for this is supported by a 2003 Federal Highway Administration (FHWA) study, which found that reducing large-truck speed by 10 mph - from 65 to 55 mph can reduce emissions of NOx. Not only does reducing the speed help enhance urban air-quality but it also increases the safety of the transportation system. When large-trucks speed, it is obviously easier for them to loose control and potentially cause harm to themselves or other drivers. This article explores the use of license-plate recognition (LPR) cameras for large-truck speed monitoring and enforcement purposes.
While reducing truck speed limits is a relatively simple act for metropolitan planning agencies, the enforcement side of it often meets with more challenges. This is the case for Tennessee Highway Patrol (THP), which has jurisdiction over Interstates 40 and 75 (I-40 and I-75), which both pass through the Knoxville metropolitan area. After the new speed law was enacted, THP found itself facing 12 million large-trucks, of which most were going faster than the 55 mph posted speed on this stretch of interstate. As can be seen in the chart "Truck Speed Comparison over Years", about 80 percent of the trucks are traveling faster than the posted speed limit of 55 mph. A complicating issue for the THP was that it was provided no annual budget or manpower increases for the purpose of enforcement. So the question is how can the new speed limit be enforced with limited manpower?
In order to determine if large-trucks are speeding along I-40, the University of Tennessee (UT) has taken advantage of the use of LPR cameras to automatically track large-trucks as they cross through the metropolitan area. Two LPR cameras would be used to capture the plate numbers (large-truck license plates are located on the front of the vehicle) of westbound trucks, accompanied with an image of each plate, as well as record the time that the plate was captured (this is a key element in determining the speed of each truck). These captured plate numbers will be matched to determine the speeds. The problem with LPR cameras is that they have a difficult time recognizing the characters on a license plate. From previous studies, LPR units rarely recognize more than 80% of the plates and often recognize less than 60%. To reconcile for this, UT applied a "matching algorithm" (procedure) to compensate for the errors made by the LPR units. With this procedure, two imperfectly read plates can be compared and one may be able to say with reasonable judgment that the plates are a match.
In order to verify this procedure, data was collected with two LPR cameras along I-40. One LPR camera was setup 1.4 miles upstream of a weigh station, which is located on the western end of the area where all trucks are required to stop for inspection, and the second camera was setup at the beginning of the weigh station. Over a 5 day period, 2,671 plates were captured and the first setup and 1,530 were captured at the second setup. Among these, a total of 787 plates were manually verified (by looking at the images captured by the first camera, writing down the characters on the license plate, and then comparing them to the set of images captured by the second camera) as identical (meaning that the plate images were captured at both setup locations). After applying the matching procedure, 787 plates were hoped to be matched.
Regarding the accuracy of the individual LPR units in the field, the first setup correctly captured 61% of the plates and the second 63%. While the accuracy of the LPR cameras was less than perfect, most of the plate characters were recognized, even if incorrectly sometimes; the use of the matching procedure increased the number of positive matches between the two setups to 97% (only about 2% were matched incorrectly). Further research is still needed to improve the reliability of the LPR cameras for the use of speed-enforcement. One of the research directions the authors are continuing on is to identify the way that the LPR equipments misread certain characters (for example, there is a relative high chance of the characters "1", "0", and "B" being misread as "I", "O", and "8", respectively), tabulate this information and use it to improve the performance of the matching procedure.
After further research, hopefully a reliable LPR speed-enforcement system, with equipment strategically located along I-40 could issue warnings or citations even while the perpetrating trucks stop on the weigh scale, with a THP officer stationed in the weigh house. This system would function in real time without the need for mailing out speeding tickets after the fact or pulling trucks over after dangerous high-speed pursuit, both of which alternatives are resource- and labor-intensive. After some time frame, truck drivers would slow down in order to reduce their chances of receiving a ticket, therefore increasing the safety of the roadway.