针对视频序列的超分辨率重建,提出了一种动态自适应滤波方法. 在最大后验概率估计和加权最小二乘

的基础上,给出视频序列超分辨率重建数学模型 深入研究了运动补偿矩阵和权值矩阵的构成和性质 详细推导了

自适应滤波器的递推公式 分析了算法的存储与计算复杂度. 仿真实验表明该算法的重建结果相当有效,相比双三

次插值和无运动补偿的单帧迭代重建,可以获得一定的PSNR 增益 与Elad 滤波方法相比,具有更小的计算量和

更强的自适应性和鲁棒性.-Super-resolution for the reconstruction of video sequence, a dynamic adaptive filter. The maximum a posteriori probability estimation and weighted least squares, based on the super-resolution reconstruction of video sequence is given mathematical model　in-depth study of the motion compensation matrix and weighted matrix composition and properties　detailed derivation of the adaptive filter recurrence formula　analysis of algorithms for the storage and computational complexity. Simulation results show that reconstruction of the algorithm very effective, compared to bicubic interpolation and non- Iterative reconstruction of single-frame motion compensation, allowed a certain PSNR gain　and Elad filtering method, the calculation has a smaller volume and greater adaptability and robustness.