摘要： 提出一种稳健的超分辨率重建算法，并在影像鉴定中加以具体运用。通过对边缘图像进行建模，构建基于洛伦兹分布的最大后验概率重建算法，并自适应地调整洛伦兹宽度参数来逐步修正迭代结果。实验结果证明，该算法能获得理想的图像画面重建效果，具有较强的稳健性。

关键词: 影像鉴定, 超分辨率, 洛伦兹分布

Abstract: This paper proposes a stable Super Resolution(SR) reconstruction algorithm to image analysis for enhancing the quality of images. It proposes Lorenztain-based Maximum A Probability(MAP) algorithm by modeling the edge image, which is iterated by the Lorentzwidth parameter adaptively to modify the reconstruction. Experimental results demonstrate validity and superiority of the algorithm, which can get ideal and stable reconstructions when applying to image forensics practices.

Key words: image analysis, Super Resolution(SR), Lorentzain distribution

中图分类号: 

• TP391