Simple, effective rate control for video distribution in heterogeneous intelligent transportation system networks

We introduce a simple video rate control strategy for reducing or eliminating decoder errors due to network errors and dropped packets on low bandwidth links in heterogeneous intelligent transportation system video networks. The proposed method is based on adaptively dropping frames prior to H.264 encoding. Performance of this strategy is assessed by evaluating three video quality metrics (PSNR, SSIM, and VQM) on actual traffic surveillance video that is transmitted across links with simulated network errors. We show that high spatial quality can be maintained at the expense of temporal resolution by reducing the effective frame rate as low as five frames per second, which is still sufficient for traffic network monitoring. Our solution is driven by several specialized constraints. Hence, while it is highly effective for intelligent transportation systems, this approach would not be suitable for more general video entertainment applications.