论文部分内容阅读
基于网络中心节点的运动矢量外推技术,提出了一种无线视频传感阵列的低复杂度多视点视频编码方法。该方法考虑到密集型视频传感阵列各视点间通信复杂、布线繁重、且位于各相机节点内的编码器由于计算能力、功耗等限制难以完成复杂的编码过程等特点,利用运动矢量外推逼近技术将大量的运动估计运算从视频编码端移到了网络中心节点,使得新编解码框架下编码器的运动估计的计算复杂度只有传统全搜索运动估计运算的0.3%,降低了系统传感阵列编码端功耗。实验结果表明,该方法的率失真性能比H.264-I帧高出4 dB以上,接近H.264-P帧编码,优于基于Wyner-Ziv理论的分布式多视点视频编码方法。
Based on the motion vector extrapolation of network center nodes, a low complexity multi-view video coding method for wireless video sensor array is proposed. The proposed method takes into account the complexity of communications between the various viewpoints of the dense video sensor array and the heavy wiring, and the encoder located in each camera node is difficult to complete the complicated encoding process due to the limitation of the computing power and the power consumption. The motion vector extrapolation The approximation technique moves a large number of motion estimation operations from the video coding end to the network center node so that the encoder’s motion estimation under the new codec frame has a computational complexity of only 0.3% of that of the traditional full search motion estimation algorithm, reducing the system sense array Coded power consumption. The experimental results show that the proposed method is more efficient than H.264-I frame by more than 4 dB, which is close to H.264-P frame coding, which is better than the distributed multi-view video coding based on Wyner-Ziv theory.