提出了一种简单、通用的基于自适应正则化技术的MAP超分辨率重建新方法,该方法可以自适应地选择正则化参数,并充分利用中间重建结果的信息,不断对其进行更新,通过迭代过程得到最终的重建影像。利用模拟序列影像和真实序列影像分别对本文方法进行了验证,结果表明,本文方法可以根据不同影像序列的特点,自适应确定相应的正则化参数,并能找到最优解,适用性较强。
A new adaptive regularized super-resolution reconstruction method for digital image sequence is introduced. In this method, the super-resolution problem is approached from a MAP estimation framework, and a technique to determine the regularization parameter adaptively is presented. The proposed method is tested on simulated image sequence and real image sequence. Experiment results indicate that this method is better than the conventional reconstruction methods on both visual judgment and objective statistical analysis.