Progressive Streaming of Video Data for Traffic Surveillance

Abstract

Traffic surveillance has been one of the essential attributes in smart city concept. Nowadays, in such applications rotating camera is preferred over static camera. Motivation behind this substitution is to reduce the cost of data transmission and Total of cost of ownership. To design an optimal and performant wireless ‘smart city area network’ for video surveillance systems, this paper focuses on some key areas, namely, transmission efficiency, lossless video data coding, data congestion, edge computing at transmission nodes. The end objective is to achieve high quality received video stream in spite of compressed data transmission. Some research initiatives in this domain are pertinent. For example, Structural Similarity Index (SSIM) based rate distortion optimization is an effective tool in enhancing the perceptual video quality in wireless environments. However, prevailing system does not consider the network congestion conditions, affecting quality of received video. Also, effect of distortion introduced by ‘channel noise’ is unattended. This motivated us to propose a new dual metric traffic control mechanism, wherein both metrics i.e. distortion and data congestion are considered. It is based on an ‘improvised SSIM’ method which incorporates the ‘Rate of allocation’ algorithm as a function. Experimental results unveil that the proposed traffic control using similarity index under noise diversity can achieve better video quality and more data throughput.