Rural Data Collection Technology - FHWA Operations - U.S. ...

More documents

Recommendations

Info

2.4 Automatic License PlateReaders (ALPR)At a Glance• Optical cameras capture images of licenseplates and software “reads” the information• Mature technology (over 30 years)• Installed above the roadway and requiredirect line-of-sight• Particularly sensitive to factors thatreduce visibility• Potential privacy issues are a concernHow It WorksOptical cameras capture images of license plates ofoncoming or receding traffic and use video imageprocessing to “read” the license plates. License platenumbers can then be matched at sensor locationsdownstream to generate travel times. Cameraimages can be stored, though this is not requiredfor travel time data generation. While ALPR is oftenimplemented as a stand-alone function, it can alsobe considered as one potential function of a machinevision system (see Section 2.5). One advantage ofALPR is that nearly all vehicles have a license platethat can potentially be observed.State of the TechnologyALPR is a relatively mature technology that has beenused for more than 30 years, though technology hasimproved and become substantially less expensivesince the earliest implementations. ALPR has beenused for tolling, law enforcement (e.g., detection ofunregistered/stolen/warranted vehicles, automatedspeed enforcement), and real-time travel time.Hardware and InstallationALPR uses cameras that operate in the visible lightspectrum. Cameras require direct line of sight tolicense plates, so they must be installed above theroadway or on the roadside in locations that minimizevisual obstructions (e.g., from surrounding traffic) andavoid off-axis angles that could reduce recognitionaccuracy. To trigger image capture, cameras can usevideo image processing to detect vehicle presencein the frame, or a separate presence-detectiontechnology can be used (e.g., inductive loop). Atnight, cameras may need additional visible or infraredillumination for adequate license plate recognition.Figure 6 shows an ALPR camera and illuminatorused in Arizona. Current systems most often useone camera per lane for license plate recognition.Implementation ConsiderationsALPR can be used for a variety of purposesindividually or simultaneously. For instance, it cantheoretically be used to detect unregistered vehicles(i.e., cross-reference license plate readings againstvehicle registration database), conduct averagespeed enforcement (i.e., detect an individual vehicle’sexcessive speed over a road segment), serve as aclosed-circuit television (CCTV) feed, and serve as atravel time data source all at the same time. Becauseit relies on a clear view of license plates, however,ALPR is particularly sensitive to any factors thatreduce visibility, such as precipitation, lens fog,line-of-sight obstructions, low ambient light, off-axisviewing, and license plates that are dirty, obstructed,missing, or have low character contrast. In a workzone implementation on a two-lane rural highway inArizona, ALPR achieved a 60 percent recognitionrate and a segment license plate matching rate of11 percent, which was considered sufficient forthis implementation (FHWA, 2004). Not all Statesrequire vehicles to have a front license plate, so ALPRmay have the highest recognition rates using rearplates. Depending on how the system is implemented,data bandwidth may be especially high for ALPR(e.g., if video or camera images are transmitted).CostsThe FHWA Knowledge Resources database providessome cost examples for ALPR. Eberline (2008)provides cost estimates for an ALPR system usedon freeways in Arizona to detect unregistered anduninsured vehicles. He estimates that the camerascost $20,000 each with an approximate installationand supporting hardware cost of $4,000 per camera.Given that one camera is required for each monitoredlane of traffic, actual costs will be higher if multilanecoverage is needed. The Texas TransportationInstitute estimated that ALPR would cost $25,000for a four-lane installation, where each lane has itsown sensor (ITS International, 2010).Rural Data Collection TechnologyFederal Highway Administration14
Figure 6ALPR Illumination Foreground) and Camera (Background)Used in a Work Zone on Arizona State Route 68Source: FHWA, 2004Privacy IssuesALPR records license plate numbers, and potentiallycamera images or video of vehicles, which canbe used to identify vehicle owners by crossreferencingmotor vehicle records. Depending on theimplementation, drivers may be identified by video orcamera images. The American Civil Liberties Unionhas expressed concerns over the potential for ALPRto be used to track individuals without a warrantand has sued Federal agencies for access to theirrecords on license plate tracking (Crockford, 2012).Agencies can minimize privacy concerns by limitingtheir access to personally identifiable informationand adhering to clearly stated policies for what canand cannot be done with ALPR data. For instance,ALPR systems can be designed to prevent operatoraccess to license plate numbers or images by deletingrecords immediately after use and providing onlyrandomly reassigned identifiers to agencies for recordkeeping purposes.Future ConsiderationsALPR is a relatively mature technology with a broadset of potential uses. Improvements in digital cameraand image processing technology are likely to leadto continued enhancements of these systems as wellas reduced costs. Perhaps the greatest potentiallimitation of ALPR in the future relates to privacyissues, as discussed above.Rural Data Collection TechnologyFederal Highway Administration15
Page 1 and 2: Travel Time on Arterials and Rural
Page 3 and 4: Technical Report Documentation Page
Page 5 and 6: List of FiguresFigure 1 RTT signs i
Page 7 and 8: List of AcronymsALPRAutomatic Licen
Page 9 and 10: Table 1 Candidate RTT TechnologiesT
Page 11 and 12: Cell Phone Signal MonitoringCell ph
Page 13 and 14: 1 Introduction1. 1 Background and O
Page 15 and 16: The second objective (experiences)
Page 17 and 18: 1.3 Organization of SynthesisReport
Page 19 and 20: Table 3 Candidate RTT TechnologiesT
Page 21 and 22: time when the vehicle is detected a
Page 23 and 24: State of the TechnologyToll tag rea
Page 25: Figure 5Typical Sensys implementati
Page 29 and 30: expand. For instance, increased ima
Page 31 and 32: measurement, which can be relativel
Page 33 and 34: approach were to be used by a trans
Page 35 and 36: 3 Implementations of Rural Travel T
Page 37 and 38: DisseminationRTI’s flagship softw
Page 39 and 40: Figure 10State of MaineVariable Spe
Page 41 and 42: supports efforts to distribute ITS
Page 43 and 44: 3.5 Various Routes in Oregon,Fronti
Page 45 and 46: 3.7 I-90, Snoqualmie Passin Washing
Page 47 and 48: 4 Best Practices for Rural Travel T
Page 49 and 50: in delayed processing in order to a
Page 51 and 52: How Long is the Data Path?It is not
Page 53 and 54: • TMC software and operators have
Page 55 and 56: ReferencesAthey Creek Consultants (
Page 57 and 58: ReferencesWashington State Departme

Rural Data Collection Technology - FHWA Operations - U.S. ...

Create successful ePaper yourself

Delete template?

Save as template?