Study On Video Enhancement Of Post-Processing

With the coming of the digital television broadcasting system and the development of high definition flat panel display (FPD), people put more emphasis on the quality of video signal. There are many processing steps before video signal coming into television viewer (TV). Information loss during the process causes the degradation of video signal. Besides, characters of FPD also impair the displaying performance. There are two kinds of degradations that impact the quality of video signal, i.e. Dynamic and Static. Dynamic means degradation appears only when objects are moving in video, such as motion blur and motion judder; while Static means degradation exists during the whole displaying process whether objects are moving or not, such as noise, edge blurring and the contrast problem. One of the feathers of video post-processing module, which connects video decoder and display device, is to solve these problems and enhance video signal. The evaluation of the performance of video post-processing mainly depends on subjective ways, therefore, there are no standards in this field and companies often develop products according to their understanding of technology and market. So the research on video post-processing is meaningful.The thesis researches on frame rate up conversion (FRC), adaptive contrast enhancement (ACE) and blocking noise reduction filter (BNR), all of which are important modules in the video post processing. These technologies enhance the quality of video directly but solve different problems. A frame rate up-conversion algorithm based on matching classification and dominant object tracking is proposed to solve the problems of motion judder and motion blur. The algorithm adds matching results analyzing and dominant object tracking modules into conventional bilateral motion estimation (BIME) algorithm and focuses on dominant motions to which human visual system is sensitive in high frame rate. The algorithm ensures blocks in dominant objects interpolated with little artifacts. To enhance video contrast automatically, a modified bilateral histogram equalization algorithm (BBHE) is proposed. It adjusts the histogram before BBHE and also processes color components. Experimental results show that the proposed method gets a good performance. According to the criterion of mean squared difference of slope (MSDS) and loop filter in H.264, a deblocking algorithm is proposed. It reduces the block noise to a great extent.