Mobile Tv Source Coding Motion Estimation Search Algorithm

Mobile TV, which has the capabilities of both broadcast television and mobile communication, is a kind of new media that appears with the development of times. Its biggest advantages are convenience and real-time feature, which makes people have the ability to receive TV programs or download news and information through mobile phones no matter where they are or what they are doing. There are a huge amount of people who use mobile phone in China, which makes china has a larger room for the development of mobile TV.In this paper, we had the research for the key technologies of mobile TV through analyzing the mobile TV system and technology programs. In this study, we mainly focused on H.264 which is the source coding standard for mobile TV and AVS which has the independent intellectual property rights. After analyzing H.264 and AVS, we concluded that both of them have the similar structure and they show the high performance in coding compression. However, the high performance costs them low computational complexity. In order to low the computational complexity, we made a detailed study on motion estimation which is the main time-consuming part in H.264 standard. At the beginning, we studied the concept of motion estimation, including blocking-matching motion estimation and several classical search algorithms, and at the same time, we presented the method about how to improve the algorithm after analyzing the disadvantages of algorithm. Through studying the characteristics of motion vector distribution and NCDS algorithm, we presented an improved algorithm named INCDS, which mainly focused on changing initial search point, early exit method, search direction and image content division to improve the performance of the algorithm and low the time that motion estimation spends.Simulation results showed that INCDS algorithm can improve the speed of motion estimation and low the search points greatly while ensuring the quality of image. In order to compare the quality of reconstruction image directly, we drew the curve chart that showed the PSNR Y value of 100 frames in the reconstruction image, from which we can see the trend of PSNR Y value of reconstruction image clearly.