Research On Multimedia Video Compression Algorithm For Wireless Sensor Network

As a new and emerging type of sensor networks, Wireless Multimedia Sensor Networks haveapplied more and more for completely monitoring the target enviroment.Different from thetraditional Wireless Sensors Networks which usually measure scalar physical phenomena, WirelessMultimedia Sensor Networks focus on the capture and processing of large multimedia informationlike audio and video stream. In the Video application, the video compression can reduce the datainformation of Wireless Sensors Networks. Nowadays, many typical Video Compression CodingStandards (like MPEG-X, H.26X) adopt motion compensation on the Encoder End. So thecomplexity of Encoder is 5 to 10 times then the Decoder that can't met the demand of WirelessMultimedia Sensor Networks which not only have a limited network resource but also the lowcomplexity and low power consumption node. Since the Distributed Video Coding adopt"Intraframe encode + Interframe decode" and excavate the dependence of video sequence on theDecoder, so the complexity of Encoder is lower and make the Encoder more brief. The distributedvideo coding is very applicable to solve the video coding issues on the Wireless Multimedia SensorNetworks.This thesis is based on the Distributed Video Coding. Aim at the issues that traditionalDistributed Video Coding can't decode the intense motion and boundary regions exactly, this thesisproposed an algorithm which based on gradient field. The algorithm utilize surface fitting toproduce the image gradient field and extracts such areas through the criteria of region of interest,then decode such areas by Expectation Maximization iteration algorithm and decode other areas byLDPC. Since the quantification and Discrete Cosine Transformation on the encoder will bring upthe block-effect and the algorithm apply Deblocking filtering on the decoded picture. Simulationresults show that: the algorithm not only can enhance the decoding effect of the intense motion andboundary regions without increase the complexity of encoder but also improve the subjective visualeffect and Peak Signal to Noise Ratio of decoded picture.