| 引用本文: | 钟山,陈洪光,沈振康.基于MAP估计的图像超分辨率重建.[J].国防科技大学学报,2003,25(4):59-62.[点击复制] |
| ZHONG Shan,CHEN Hongguang,SHEN Zhenkang.MAP-Based Image Superresolution Reconstruction[J].Journal of National University of Defense Technology,2003,25(4):59-62[点击复制] |
|
| |
|
|
| 本文已被：浏览 6415次 下载 6157次 |
| 基于MAP估计的图像超分辨率重建 |
| 钟山, 陈洪光, 沈振康 |
|
(国防科技大学 ATR重点实验室，湖南 长沙 410073)
|
| 摘要: |
| 从一组低分辨率图像中重建一幅高分辨率图像，给出基于MAP估计的近似算法。描述从场景空间到观测图像的线性映射关系，介绍了核心重建算法；算法适合观测条件相似的情况。在贝叶斯方法中，使用Laplace算子作估计图像的领域平滑相关性假设。最后的实验表明了算法在改善空间分辨率上的显著效果。 |
| 关键词: 超分辨率 MAP Laplace 算法 图像 |
| DOI： |
| 投稿日期：2003-02-24 |
| 基金项目: |
|
| MAP-Based Image Superresolution Reconstruction |
| ZHONG Shan, CHEN Hongguang, SHEN Zhenkang |
|
(ATR Lab, National Univ. of Defense Technology, Changsha 410073, China)
|
| Abstract: |
| We describe a MAP-based method for constructing a high resolution image from the noisy undersampled multiframes. Using the aliasing relationship between the high resolution image and the undersampled frames, we develop the algorithms for the reconstruction problem, appropriate for the case where all images are taken under similar conditions. The Bayesian approach uses Laplacian operator to model a neighbor correlation on estimated image. The results of experiments show dramatic improvement in the spatial resolution. |
| Keywords: superresolution MAP Laplace algorithm image |
|
|
