Abstract
Transportation agencies have invested significantly in extensive closed-circuit television (CCTV) systems to monitor freeways in urban areas. While thes systems have proven to be very effective in supporting incident management, they do not support the collection of quantitative measures of traffic conditions. Instead, they simply provide images that must be interpreted by trained operators. While there are several video image vehicle detection systems (VIVDS) on the market that have the capability to automatically derive traffic measures fro video imagery, these systems require the installation of fixed-position cameras. Thus, they have not been integrated with the existing moveable CCTV cameras. VIVDS camera positioning and calibration challenges were addressed and a prototype machine-vision system was developed that successfully integrated existing moveable CCTV cameras with VIVDS. Results of testing the prototype are presentedindicating that when the camera’s initial zoom level was kept between ×1 and ×1.5, the camera consistently could be returned to its original position with a repositioning accuracy of less than 0.03 to 0.1 regardless of the camera’s displaced pan, tilt, or zoom settings at the time of repositioning. This level of positional accuracy when combined with a VIVDS resulted in vehicle count errors of less than 1%.
Get full access to this article
View all access options for this article.