Abstract
Wireless Video Sensor Networks (WVSNs) - a type of WSNs - comprise of sensor nodes that can capture, process and communicate video frames. The battery powered sensor nodes have limited hardware resources while video processing and communication are resource intensive tasks i.e., require high-end processors, large memory and bandwidth. Video encoding is a popular method used to reduce the communication overhead but being an inherently complex process it results in higher computational energy-drain on video sensor nodes. This establishes an interesting computation-communication tradeoff for energy efficient video communication (encoding and transmission) in WVSNs. In this paper, we study this computation-communication tradeoff under Intel-imote2 based single-hop and multi-hop video sensor networks testbed by empirically evaluating selected implementations of the MPEG-4 (Part 2) and H. 264/AVC encoders. The analysis has been carried out to characterize the performance of encoders in terms of energy efficiency, compression efficiency and video distortion. The experimental results show that in single-hop WVSNs, MPEG-4 is energy efficient over H. 264 whilst in multi-hop WVSNs, H. 264 is energy efficient over MPEG-4.