Extreme CCTV License Plate Reader Field Tests Successful for Florida DOT

May 22, 2006
REG license plate capture cameras tested for travel time applications
VANCOUVER, BRITISH COLUMBIA - Extreme CCTV International Inc., a subsidiary of Extreme CCTV Inc. (TSX:EXC), announces that the results of a field test conducted for the Florida Department of Transportation (FDOT), demonstrate the effectiveness of REG license plate capture cameras for travel time applications in Intelligent Transportation Systems (ITS).

The test, conducted for FDOT by PBS&J, a nationwide leader in transportation consulting, investigated the viability of license plate reading (LPR) technology for use in travel time data collection. The test took place over three weeks, each week at a different location collecting five days (Monday to Friday) of traffic data over multiple lanes. Southern Traffic Services of Pensacola, FL provided all services associated with project installation.

Although REG is currently being successfully applied for travel time ITS projects in Europe, license plate capture is not widely used in North America due to the technical challenges presented by effective imaging of North American plates. Lack of design standardization, variations in reflective properties and the presence of logos, emblems and other insignia all pose significant technical challenges for any LPR system in North America.

Conducted in Tallahassee, the test involved some of the most difficult LPR conditions that could be encountered in the field. Florida license plates are constructed of a particularly low-contrast material, making the capture of a usable image especially difficult. To further complicate the task, Florida plates come in many different visual designs, resulting in poor recognition rates in conventional LPR systems. Additionally, several of the test sites were located at major roadways where the speed of the vehicles averaged approximately 70 miles per hour. All these factors combined made the Florida test technically demanding.

Once captured, the specialized LPR software utilizes a video trigger to begin a sequence of several "snapshots" of the moving vehicles as they pass through the field of view of the REG® camera. The software then determines the best image from the set of snapshot and uses it to perform an optical character recognition (OCR). The OCR functions to transcribe the image of the license plate into an alphanumeric text string for further processing. For the field test, Extreme CCTV selected an OCR engine that utilizes template and shape matching to ensure high accuracy in the results. After the OCR, the software stores the license plate as plain text and associates the license plate image from which it was derived.

Based on their analysis of the collected test data, PBS&J determined that significant numbers of license plate images were collected at field test LPR sensor locations. Using the LPR sensor data, PBS&J subsequently calculated travel times for the distances between pairs of LPR sensors and plotted these travel times against the time of day.

In their conclusions, PBS&J indicated that a REG-based LPR system has the potential to provide a viable means of collecting travel time data. They also suggest that even in situations where only a single lane of LPR-based data collection infrastructure is available, it is still possible for a system to produce usable travel time data under the right circumstances (high traffic volumes and read rates), especially along limited access roads. PBS&J plans to use the results of this test to assist FDOT in enhancing the accuracy, timeliness and reliability of real-time traffic data collected on Florida's roadways.