摘要
实现序列图像的超分辨率重建,需要利用同一场景的多幅低分辨率图像之间的相对运动信息,并将它们融合到单幅高分辨率图像中,以有效的去除低分辨率图像中的模糊和噪声。本文提出首先分析序列图像结构、纹理等多维特征的不同特性和作用,利用分解得到的多维特征分别采用凸集投影(POCS)、范例学习等具有针对性的重建方法进行图像放大,在有效融合多维特征重建图像的基础上,实现序列图像的多维特征超分辨率重建。
For implementing sequence image super-resolution reconstruction, need to use more of the same scene ot relative movement between the low-resolution image information, and put them into a single high-resolution image, to effectively remove the blur and noise in the low-resolution image.In this paper, first of all, analysis the different characteristics between the sequence image structure and the texture, using the targeted image super-resolution reconstruction methods POCS and examples-study for the decomposed multi-dimensional characteristics, implemented sequence image super-resolution reconstruction based on the effective fusion for the nmlti-dilnensional reconstruction images. Keywords: Super-Resolution Reconstruction; Sequence Image; Multi-dimensional; POCS; Example-Study
出处
《网络安全技术与应用》
2014年第3期200-200,203,共2页
Network Security Technology & Application
关键词
超分辨率重建
序列图像
多维特征
凸集投影
范例学习
Super-Resolution Reconstruction
Sequence Image
Multi-dimensional
POCS
Example-Study