摘要
在提高图像分辨率过程中,选择合适的成像模型是获得高分辨率图像的一个关键因素。提出一种基于最大后验概率、采用Gibbs成像模型,并利用多帧序列图像重构高分辨率图像的方法,充分利用序列图像之间的累加信息,从而快速地获得较其他方法更优的高分辨率图像。
During the image restoration, choosing a proper model that describes the image is the key of restoration. A MAP framework based on Gibbs to extract a single super resolution image from image sequences is proposed. Additional novel observation data can be obtained and the feasible solution space is constrained with apriori assumptions on the form of the solution. Experimental results are provided to illustrate the performance of the algorithm that is better and faster than others.
出处
《红外与激光工程》
EI
CSCD
北大核心
2003年第6期613-616,共4页
Infrared and Laser Engineering
关键词
高分辨率
运动向量
序列图像
最大后验概率
High resolution
Motion vector
Image sequences
Maximum a Posteriori(MAP)
