基于数学形态学的视频帧间压缩算法研究与实现

Research and Implementation of the Video P-frame Compression Algorithm Based on Mathematical Morphology

常用视频压缩算法DCT存在方块效应和"飞蚊噪声",在低速率传输下更为明显,难以满足视频传输中带宽窄、质量高的要求。基于数学形态学的视频图像帧间编码技术,是运用帧间运动补偿的编码算法,该算法通过复合形态变换,用结构元素对宏块中的不规则小块进行开、闭运算,选取并滤出不规则小块,再对不规则小块的大小和运动矢量进行运动估计编码,采用绝对差分和找出匹配块,实现帧间预测。通过仿真可达到比较精细的补偿,提高了编码效率,在相同压缩比下可提高视频质量,弥补了DCT算法在图像传输中的不足,改善了图像传输效果,降低了运算复杂度,对网络视频图像的编解码及传输具有一定的实用意义。

DCT commonly used algorithms in video compression exist the square effect and ＂ mosquito＇s noise flies ＂,especially under the low speed.Neither take up the narrow bandwidth in the satisfied video transmission nor transmit the high request in qualities.And mathematics morphology can remedy the deficiency in narrow bandwidth.To the P-frame,this thesis adopt a kind of newtype algorithm of motion compensation based on the transform of the mathematical morphology.Through compound morphological transform,the structure elements carry out open and close operation with irregular minuteness blocks to filter them,then,motion estimating and coding with size and sport vector of them,adopting SAD to find out matching blocks.In this way,we can achieve the purpose of frame prediction.Through computer emulation,the algorithm achieve fine compensation comparatively and improving the compression rate.Under the same compression rate,it can improve video quantity,to make up for the shortcomings of DCT algorithm in the image transmission,improve transmission effect and reduced operation complexity.The algorithm has certain practicability in video coding or encoding for internet.