Research On Vision Sensor Nodes Based On Parallel Computing Of Distributed Video Coding

Wireless visual sensor networks(WVSNs)provide rich video information and enable more sophisticated content monitoring.In recent years,it has developed rapidly,but video signal has a large amount of data,high coding complexity,and high requirements for nodes of WVSNs,which has become a bottleneck restricting the development of WVSNs.Distributed video coding(DVC)shifts the coding complexity from the encoding end to the decoding end,which effectively reduces the energy consumption of the sensor node,and is more suitable for WVSNs.However,if the entire coding task is concentrated to a single sensor node,the amount of computation will still cause the corresponding node to consume too much energy and affect its life cycle.The paper proposes a parallel computing method for DVC,which divides the encoding structure into several levels.The coding tasks in each level are decomposed into multiple computational sub-tasks that can be calculated in parallel,and are shared by the remaining member nodes in the cluster to balance the energy consumption of the network.Simulation experiments show that parallel computing effectively shortens the coding time of DVC.Compared with DVC performed by a single node,the higher the number of video frames,the higher the coding efficiency of parallel computation.The parallel computing method of DVC provides a new idea for WVSNs,which is of great significance to its development.In the paper,a wireless visual sensor node based on FPGA is designed,and the detailed circuit design and selection for hardware modules such as microprocessor module,power module,image acquisition module,memory module and wireless transceiver module are carried out.The logic design of various software modules including global clock,distributed video coding,and wireless transmission are completed by Verilog language.Experimental results show that the designed node can complete functions including video acquisition,compression coding and Wi-Fi transmission,which provide physical platform support for the research and application of the parallel computing of DVC in WVSNs.