Video CODEC with adaptive frame rate control for intelligent transportation system applications

Abstract

Video cameras are one of the important types of devices in Intelligent Transportation Systems (ITS). The camera images are practical, widely deployable and beneficial for traffic management and congestion control. The advent of image processing has established several applications based on ITS camera images, including vehicle detection, weather monitoring, smart work zones, etc. Unlike digital video entertainment applications, the camera images in ITS applications require high video image quality but usually not a high video frame rate. Traditional block-based video compression standards, which were developed primarily with the video entertainment industry in mind, are dependent on adaptive rate control algorithms to control the video quality and the video frame rate. Modern rate control algorithms range from simple frame skipping to complicated adaptive algorithms based on optimal rate-distortion theory. In this dissertation, I presented an innovative video frame rate control scheme based on adaptive frame dropping. Video transmission schemes were also discussed and a new strategy to reduce the video traffic on the network was presented. Experimental results in a variety of network scenarios shown that the proposed technique could improve video quality in both the temporal and spatial dimensions, as quantified by standard video metrics (up to 6 percent of PSNR, 5 percent of SSIM, and 10 percent VQM compared to the original video). Another benefit of the proposed technique is that video traffic and network congestion are generally reduced. Both FPGA and embedded Linux implementations are considered for video encoder development.