Design Of Video Motion Estimation And Noise Suppression Filtering Components

MPEG is a series of high quality international video and audio compression standards; they have wide variety of applications in such areas as video storage, video broadcasting and video stream media transmission. According to the principle of MPEG-2 and MPEG-4, it is possible to design a general video coding platform for SoC design, in which noise suppression and video motion estimation are two important components. Because of the complexity of algorithm and the real time requirement of application, these two components are realized in hardware.The noise suppression component removes high frequency noise from the original image by hardware filtering, making the image smoother. In this paper, the type of image noise is analyzed, median filter is chosen according to high frequency impulse noise, the architecture of a high performance adapt median filter is proposedMotion estimation component is the most important and complex component in video encoder. Algorithms for motion can be divided into two types: full search and fast search. Full search algorithm has high precision; the corresponding architecture has clear data path and simple control structure. But according to its huge calculation complexity, the according architecture has long calculation time and large hardware overhead. Fast search algorithms can largely reduce search time with effects on precision. In this paper, one of the best algorithms, three-step hierarchical search algorithm is chosen, and simple and high performance architecture is proposed.