面向嵌入式微处理器的高性能视频编解码算法OA北大核心CSTPCD

In allusion to the situation that the embedded microprocessor is often limited by resources in the video codec of high definition and ultra high definition video,which leads to the blurring of the output video picture,a high-performance video coding algorithm for embedded microprocessor is proposed.The minimum average absolute error is used as the matching criterion index to pre-judge the distortion rate loss of high-performance video,calculate the total energy of the maximum AC coefficient in the image,and obtain high-detail segmentation and motion sensing video.The weights of adjacent pixels in the video coding are obtained by means of the edge graph in the space-time domain,and the probability that the superpixel is the foreground region is obtained by means of the distance between pixels,and the target foreground region is encoded.The minimum norm is used to optimize the constraint of coding video,the Lagrange multiplier equivalent transformation is used to encode video,and the video decoding correction factor is calculated according to the training redundancy dictionary to complete the final video codc.The experimental results show that the proposed method can effectively complete high performance video codec and ensure video quality,and the video encoding rate is always 31.3 dB,which is relatively stable.