Research On Low Complexity Distributed Arithmetic Coding And Its Application In Intelligent Transportation

As an important part of Intelligent Transportation Systems(ITS),surveillance cameras generate huge amounts of multimedia data every day.In addition,with the improvement of the construction level of intelligent transportation and the continuous addition of new vehicles such as intelligent networked vehicles,ITS will also face the explosive growth of data in the future.The data transmission must go through the encoding step.The performance of the encoding technology will directly affect the efficiency and reliability of data transmission.Especially in the face of massive multimedia data,more rapid and efficient compression coding technology is more important.Therefore,research on data coding technology is very necessary.In view of the large amount of data generated by cameras and strong correlation between videos in the dense urban monitoring scene,this paper proposes to apply Distributed Arithmetic Coding(DAC)to the above intelligent traffic scene,which provides a solution to the transmission problem of massive strongly correlated multimedia data.As an efficient data compression scheme,distributed arithmetic coding has achieved certain research results.Based on the existing research,this paper conducts a more in-depth study on distributed arithmetic coding: taking binary distributed arithmetic coding as the research object,the low-complexity approximation methods for calculating its initial spectrum are proposed,and its decoding error probability is calculated.In order to make the application range of distributed arithmetic coding more extensive,the research is further extended to Q-ary conditions,and the symbol-interval mapping rules and coding and decoding principles of Q-ary distributed arithmetic codes are analyzed in detail.The corresponding codecs are designed and used for actual traffic video compression.In this paper,the proposed methods and related applications are verified by simulation.The main works of this paper are as follows:(1)The improvements of initial spectrum calculation method of binary distributed arithmetic coding are proposed.As a theoretical analysis tool,the initial spectrum can effectively improve the coding efficiency of distributed arithmetic coding,so it has important application value.At present,the calculation of the initial spectrum is mainly realized by a numerical algorithm.However,after in-depth analysis,it is found that the calculation result of this algorithm does not fully converge to the real initial spectrum,and its computational complexity is high.To solve the above problems,this paper from different angles,respectively put forward two kinds of improvement methods.Aiming at the problem that the original numerical algorithm does not converge to the real initial spectrum,after analyzing the causes of error in detail,a fair numerical algorithm is proposed to improve it.To solve the problem of high computational complexity of the original numerical algorithm,an improved method of low complexity approximation is proposed within the range of code rate interval R∈(0,0.5].According to the actual situation,the above bit rate interval is further divided into three sub-intervals,and the corresponding initial spectral low complexity approximation method is proposed for different sub-intervals.By designing experiments,the two improved methods are compared with the original numerical algorithm from different angles.The experimental results fully prove that the two improvement methods are correct and effective.(2)A theoretical calculation method for the decoding error probability of binary distributed arithmetic coding is proposed.At present,the calculation of the decoding error probability is obtained by statistics after the source is encoded and decoded.However,at this stage,there is no method for calculating the decoding error probability from the perspective of theoretical analysis.Aiming at this research gap,this paper proposes a theoretical method for calculating the decoding error probability of binary distributed arithmetic coding by using "Coexisting Interval" as an analysis tool.First,start the study with the simplest case: all but the last symbol of each source block are known at the decoder,and then further analyze the more complex case: all but the last two symbols of each source block are known at the decoder.The analysis process of calculating the decoding error probability in the above two cases and the final theoretical calculation results are given in detail;finally,the analysis ideas for calculating the decoding error probability in a general case are summarized.In the experimental section,the corresponding codec is designed to obtain the actual statistical decoding error probability under different experimental conditions.By comparing the statistical results with the theoretical calculation results proposed in this paper,the correctness of the proposed method is verified.(3)Extend the research on distributed arithmetic coding from binary to Q-ary conditions,proposes a new design scheme of Q-ary distributed arithmetic coding,and applies it to traffic video compression.To make the application of distributed arithmetic coding more extensive,this paper focuses on the research on Q-ary distributed arithmetic coding.Firstly,the symbol-interval mapping rule under the condition of Q-ary is designed,and the basic principle of encoding and decoding of distributed arithmetic coding under this condition is introduced in detail;secondly,different implementation schemes are designed according to the principle of encoding and decoding,and the specific process of video compression using Q-ary distributed arithmetic coding is given.In the experimental section,the designed codec schemes are applied to the actual traffic video compression,and its application prospect in the dense urban traffic monitoring scene is pointed out.By comparing the compressed result with the original video frame,the actual compression effect of different design schemes is verified.At the same time,the correctness and feasibility of applying Q-ary distributed arithmetic coding to traffic video compression are proved.