Translation Model Based On Weighted Frame Of Video Signal Processing

With the rapid economic development of the information society, the requirement for the quality of video signals is becoming higher and higher. These promoted that multimedia (especially video) industry and their technology are booming. On the military, the video technology is used in missile guidance, tanks and warships simulation training. In addition, it is widely used in traffic monitoring, fingerprint identification, face identification, etc.However, the video signal may be affected by blur, de-sampling and noise in the process of formation and transmission. In order to meet the high-quality requirement of visual, it is necessary to carry out a series of processing for video signal. Motion estimation is mostly used in traditional video signal processing technology, but its computation is large and needs higher accuracy. In order to overcome these drawbacks, we can use other algorithms to replace the motion estimation, and this is a wise choice. In this paper, we put forward weighted inter-frame translation model based on the inter-frame model, use it to instead the motion estimation algorithm, and overcome some of the drawbacks of the motion estimation.In this paper, we mostly carry on the following works for the processing of the video signal based on the weighted inter-frame translation model:1. Combining the inter-frame model and non-local average algorithm, we put forward weighted inter-frame translation model, use it to instead of motion estimation in video signal processing, and apply this model to modify the inter-frame equation of the video signal.2. Under the framework of Kalman, we use the weighted inter-frame translation model, combine the theory of linear minimum variance fusion and the observation equation for adding noise equation, get a new de-noising algorithm, and finally use non-local average proposed algorithm to post-process. Simulation results show the peak signal-to-noise ratio (PSNR) of algorithm by0.6-1.6dB. The comparative experiment is non-local average algorithm de-noising of video signal.3. The observation equation is the processing of down-sampling and adding noise for original video. Use a similar processing as the video de-noising method to get a new interpolation algorithm, and finally we use non-local average algorithm to conduct post-processing. Comparative with experiment that the video signal is first dealt with the nearest neighbor interpolation, and then is dealt with the non-local averaging algorithm. Through the simulation experiments, the PSNR of the proposed algorithm in this paper to restore the video signal improved by0.7-1.7dB. 4. Applying proposed model to super resolution reconstruction, we get a better algorithm. Comparatived with the experiment that the super resolution reconstruction video signal is first dealt with the nearest neighbor interpolation, then dealt with the non-local averaging algorithm, and finally dealt with Total-Variation regularization to conduct post-processing, the new algorithm has an advantage. Simulation results show that the PSNR of the proposed algorithm in this paper to restore the video signal improved by0.1-1.1dB.