Research On Energy Saving Technology Of Wireless Video Sensor Networks

Wireless video sensor network is a group of computing, storage and communicationcapabilities of the video sensor node consists of a distributed sensing network. Comparedwith the traditional WSN, the information of the wireless video sensor network video isgreater,the features that hardware resources limited will be more prominent,and so meetcertain performance requirements under the premise of the use of appropriate energy-savingstrategies to reduce node power consumption as much as possible, to save costs andeffectively complete the information collection task has very important significance.This article does the research mainly from the data processing aspects ofenergy-saving, adopting the low complexity of distributed video coding framework andimproving it based on the framework.The main work of thesis and the research results are as follows:1. The LBT transformation is applied in distributed video coding, and its algorithm isimproved in the transform domain coding system based on the DCT transform. Inaccordance with the principle of integer lifting and DCT transform, the LBT transformalgorithm for integer lifting is improved. The improved algorithm can avoid large amountsof computation of floating point addition and multiplication, and require only a smallamount of shift and integer addition operation.Thus it’s more simple to realize. Theexperimental results show that, compared with DCT transform the calculation is greatlyreduced and performance has been a certain degree of upgrading.2. Based on the research of distributed video coding technology, the IRA code is usedin distributed video coding algorithm.The Characteristics of IRA code is applied todistributed video coding which is simial to Tubro code and LDPC code.The IRA code hasthe low complexity which is simial to Tubro code in the encoder side. In the Decoder sidethe IRA code can be used with the same iterative decoding algorithm of the LDPC codewhich not only can reduce the complexity of the decoding side but also can reduce the errorlevel layer.Through the simulation we conclude the IRA code and LDPC code has excellentperformance close to the code length is longer and better than LDPC codes, but theencoding complexity is fromO (k2)into O (k).3. The traditional Key frame frame coding zero coefficient blocks decision algorithmis improved. The algorithm uses the predicted current coding block adjacent to the block zero coefficient blocks of information to flexibly adjust the judgment threshold zerocoefficient blocks, thereby improving the zero value judgment rate. The improved algorithmavoided zero coefficient blocks of transform and quantization process, comparisons with theexisting algorithms, which could obviously improve the zero block coefficient blockdecision rate, but also saved the encoding time, which not only improved the coding speed,and also satisfyed the energy requirements of wireless video sensor networks.