Multi-Lane Detection and Road Traffic Congestion Classification for Intelligent Transportation System

Full-Body Occlusion Handling and Density Analysis in Traffic Video-Surveillance Systems

“…In the application of vision-based traffic surveillance tools, many methods have been implemented, including adaptive median filters (3), color median (4), frame differentiation (5), mixture of Gaussians (6, 7), wavelet differentiation (8), kernel-based density estimation (9, 10) and sigma-delta filters (11). Also, mixture of Gaussians is implemented with shadow elimination based on color reflectance and gradient feature (7). Recently, a sigma-delta filter was proposed to continuously recognize a stopped vehicle as the foreground object and reduce the computational cost (11).…”

“…The true occluded vehicles typically seem to reduce in scaling factor gradually and steadily while retaining their average speed fairly constant. Thus, additional processing was added to sort out true occlusion by reinforcing the algorithm with acceleration data: a a t (7) avg, th < where a avg,t is the average acceleration of a vehicle at a particular time t and a th is the threshold acceleration of 43 ft/s 2 . The acceleration data are highly unreliable without additional smoothing because of varying frame rates; however, they serve effectively in comparing the result produced from the first occlusion handling based on the scale factor.…”