Automated Plate Recognition and Truck Trip Tracking
-
2019-06-01
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
Edition:August 2016 to June 2019
-
Corporate Publisher:
-
Abstract:This study sought to apply automated license plate recognition (ALPR) technology to track trucks trips. ALPR does not work perfectly in the U.S. because of the thousands of different designs, colors, shapes, fonts, etc. of license plate from different states. To overcome this, a class of machine learning algorithms were developed to help track trucks by matching license plates read, correctly or incorrectly, by ALPR devices. While these unsupervised machine learning algorithms worked great for short distance (<10 miles) scenarios, they have never been tested for long distance scenarios, which was the main challenge of this study. Three sets of field studies were conducted at strategically selected Interstate sites in Tennessee using mobile ALPR stations. The first study tracked trucks on I-75 from Georgia to Kentucky and to Virginia via I-81. The second study tracked trucks from Georgia and Alabama to Kentucky via I-24 and I-65. The third study tracked westward trucks through Nashville via I-40 and around Nashville via I-840. The tracking distance was between 50 and 250 miles. In general, the total matching percentage ranged from 14% to 48%. This is common and largely due to spatial and temporal leakages between stations far apart.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
File Type:
Supporting Files
-
No Additional Files
More +