MAP和POCS算法实现超分辨率图像的重建

在目前超分辨率图像融合算法中，凸集投影法（Projections onto Convex Scts，POCS）和最大后验概率（Maximum A Posterior，MAP）是最为成功的两种。但它们都存在着自身无法克服的局限性。在充分分析了两者的优、缺点的基础上，把凸约束结合到MAP算法里面，使得图像融合结果有着它们两者的优良属性。实验结果证明了算法的有效性。

Reconstruction of Super-resolution Image using MAP and POCS Algorithms

Projections onto Convex Sets(POCS) and Maximum A Posterior (MAP) are now two of the most successful algorithms in super-resolution(SR) image fusion. However, they both have their own insuperable limitations. Based on the comprehensive analyses of their advantages and disadvantages. It addes the convex constraints to the MAP method and finds excellent properties of the both algorithms in fusion results. The following experimental results prove the effectiveness of the proposed algorithm.