01661038

Component

Vehicle classification information is vital to almost all types of transportation engineering and management applications such as pavement design, signal timing and many other safety applications. Vehicular length-based vehicle classification methods are widely applied but the outcome is often coarse and may not meet requirements for finer scale traffic engineering. Axle-based vehicle classification data sources often confined to the locations where heavy truck volumes are observed. Computer vision-based approach is very flexible when it comes to flexibility of deployment as video data is increasing becomes available. This paper presents a “hybrid” vehicle classification approach combining the length and axle-based vehicle classification with improved accuracy. An image processing-based vehicle classification system: CombinedLength and Axle vehicle classification Sy Stem (CLASS) tool is developed to automatically measure the length and number of axles extracted from ground-truth videos. The preliminary test on detecting and extracting the combined vehicle length and axle parameters as defined by the FHWA classification scheme appears to be successful. The developed CLASS provides an additional vehicle classification data source at a lower cost and can be applied at locations where traffic surveillance video is available.

This paper was sponsored by TRB committee AT055 Standing Committee on Truck Size and Weight.

18-06319

English

5p

2018

Transportation Research Board 97th Annual Meeting

Digital/other

References

Axles; Image processing; Traffic surveillance; Trucks; Vehicle classification; Vehicle length; Video

Highways; Planning and Forecasting; Vehicles and Equipment

Transportation Research Board Annual Meeting 2018 Paper #18-06319

TRIS, TRB, ATRI

Jan 8 2018 11:38AM