Research On Video Intra Prediction Coding Technology Based On VVC

In order to meet people's increasing demand for high-quality and low-bit-rate video,it is an important approach to study a new generation of video coding standards to substitute the new generation of high efficiency video coding(HEVC).Versatile Video Coding(VVC),as a new generation of multifunctional video coding standard,has a wider range of application scenarios than HEVC.In order to improve the coding quality,VVC adds a variety of core technologies to each module.However,the new technology will cause VVC to have problems such as high coding calculation complexity and long coding time.The purpose of this topic is to increase the speed of VVC coding in order to broaden the field of VVC applications.VVC still uses a hybrid coding framework.This paper mainly studies the intra-frame coding technology in the framework.Because of the existing algorithm in the blocks of VVC frame segmentation not make full use of video intra spatial information.However,in the decision-making of VVC model complexity is still high,and there is space for optimization.In response to the above problems,this paper make the overall study on the coding block division and mode selection tasks of the VVC coding standard.The main work is as follows:(1)The advancement process of video coding and the current situation of technical research are generalized,and the coding framework of VVC and the corresponding coding principles are described.Detailed analysis coding block dividing process and intra mode prediction coding,which lays a framework foundation for the in-depth study of this thesis.(2)In coding block division,in order to decrease the computational complexity of VVC intra prediction coding unit division,a fast division method of coding blocks based on the similarity of gradient amplitude is proposed.First,calculate the average gradient amplitude similarity deviation of the four sub-coding units in the lower layer of the current coding unit to determine whether the current coding unit is divided into quadtrees or not.Secondly,when the above division conditions are not met,the variance of the sub-block pixel variance of the multi-type tree division is obtained by traversal,and the best division method in the multi-type tree is selected based on this information.The six fixed types of division in the original division method are avoided.(3)In the mode selection,based on the statistical analysis of the rules of the coding block mode decision process,a rapid intra prediction mode selection algorithm based on gradient direction information is proposed.Using the texture pixel characteristics of the image block,the 67 modes in the VVC frame are divided into four-direction mode sets,and the current block texture direction value is judged to classify them,reducing the number of modes entering the VVC rough selection stage,thereby increasing decide the speed.The experimental results verify the advantages and disadvantages of the proposed algorithm.Compared with the original algorithm of standard test model 7.0 encoder,the algorithm proposed in this paper reduces the time consumption by 57.91% while ensuring the image quality is basically unchanged when encoding the full frame of the video sequence.And the bit rate is only increased by 1.79%.