Energy minimization for portable video devices with a hardware H.264/AVC encoder based on energy-rate-distortion optimization

Abstract

In mobile surveillance camera systems powered by battery, reducing the encoder’s compression energy consumption is one of the most critical issues to prolong its lifetime. Over the encoder’s rate-distortion optimiza-tion for compression efficiency, energy-rate-distortion (E-R-D) optimization is studied to minimize the energy consumption. However, almost of proposed E-R-D optimization methods are based on a software codec. It is not suitable for real implemented mobile camera systems for the reason why the energy consumption is too large and the encoding frame rates is not enough. In this paper, we propose the E-R-D model for the hardware codec which is restricted in power scalability. For this purpose, we develop the gate-level simulation framework to measure energy consumption with its switching activity. Within the results, we find out that GOP size can be the power scalable parameter. From the E-R-D model, we also propose the energy minimizing algorithm for portable video camera sensor with dynamic controlling of GOP size and QP. Our experimental results show that the proposed algorithm has up to 34.38% of energy consumption adopted to sequence, rate and distortion.