面向无线多媒体传感器网络的能量受限高效视频感知编码研究

With the development of the Wireless Multimedia Sensor Networks, the research of the power constrained video compression system has become the important problem. Rate-distortion theory comes under the umbrella of source coding or compression, which is concerned with the task of maximally stripping redundancy from a source, subject to fidelity criterion. In the project the rate-distortion region is analyzed for multiteminal video coding system. Moreover, the inner-bound and the outer-bound of the rate-distortion region is determined. According to the rate-distortion function, the power efficient video compressive sensing coding system is designed. Considering the relationship of power, rate and distortion, the power-rate-distortion model is proposed to optimize the coding system. In order to change the constrained optimility problem to the unconstrained problem, KKT optimality conditions should be satisfied. Thus the optimal points satisfying the KKT conditions are solved. Finally, the power-rate-distortion optimized video compressive sensing coding system is proposed. The system has the good balance in the power, rate and distortion. It is well-suited to the Wireless Multimedia Sensor Networks. In addition, the research helps to the optimal coding in Wireless Multimedia Sensor Networks.

随着无线多媒体传感器网络的飞速发展，能量受限的高效视频压缩编码成为重要的研究内容。本申请通过研究能量受限的视频压缩编码系统的率失真性能，推导其率失真区域及编码方法。同时通过分析编码器中能耗、码率和失真的相互制约关系，建立能量-率失真模型。在此基础上设计实现面向无线多媒体传感器网络的能量受限的视频感知编码系统。通过对优化的KKT条件进行求解，将求解结果用于上述编码系统，设计了能量受限的视频感知编码系统的能量-率失真优化算法。本申请的研究成果有助于解决多媒体传感器网络中视频的最优编码问题，均衡压缩效率与能量消耗量，对复杂网络下多媒体通信的发展，具有重要的理论和实践指导意义。

在低成本的多媒体传感器网络中实时高质量的传输压缩编码数据成为重要的研究内容。本课题基于压缩感知理论研究了高效的感知编码技术，分析了压缩感知编码系统的率失真性能，研究了能量受限的压缩感知编码系统中率失真优化的KKT求解问题；分析能量、码率和失真之间的相互制约关系，建立能量-率失真模型，根据能量-率失真模型的表达式，设计能量受限的能量-率失真最优的压缩感知编码系统。上述算法的有效性通过实验进行了验证。首先验证了所提模型的正确性，在此基础上测试了能量受限的率失真优化算法性能。实验结果表明该算法能根据给定能耗和码率确定最佳的参数配置，获得最佳的图像质量。