Algorithms for Estimating Mean Vehicle Speed Using Uncalibrated Traffic Management Cameras
-
2003-10-01
Details:
-
Creators:
-
Corporate Creators:
-
Corporate Contributors:
-
Subject/TRT Terms:
-
Publication/ Report Number:
-
Resource Type:
-
Geographical Coverage:
-
TRIS Online Accession Number:968589
-
Corporate Publisher:
-
Abstract:This report documents the second project, in a series of three research projects funded by the Washington State Department of Transportation (WSDOT), that will enable already deployed, uncalibrated CCTV cameras to be used as traffic speed sensors. The principle traffic speed sensors currently deployed by WSDOT are inductance loops; however, in some locations it is impractical or too expensive to install loops. In addition, a large number of un-calibrated cameras are already in place and being used by the traffic management operators to qualitatively assess traffic both on the freeway and on arterials. These projects will leverage the existing cameras to provide a quantitative measurement of traffic speed similar to that which can be obtained using loops and suitable for use in the Traffic Management System (TMS) without installing loops in the roadway. The implementation of this research will culminate with software that creates an automated system compatible with the existing TMS. This system will leverage the existing camera investment to create a new set of speed sensors that increases the geographic extent of the TMS's quantitative surveillance capabilities. In the second phase, reported on here, roadway features are used to augment the camera calibration. This overcomes the occlusion problem, or apparent blending together of small vehicles as seen in the far field of the camera images, that existed in the first phase. Activity maps, fog lines, and vanishing points are a few of the additional features used, and the details of these algorithms are described in this report. These results have also been peer reviewed and published.
-
Format:
-
Funding:
-
Collection(s):
-
Main Document Checksum:
-
Download URL:
-
File Type: