Academic Journals Database
Disseminating quality controlled scientific knowledge

图像超分辨重建技术及其应用 The Technique and Application of Image Super-Resolution Reconstruction

ADD TO MY LIST
 
Author(s): 于璐璐 | 邹华

Journal: Optoelectronics
ISSN 2164-5450

Volume: 03;
Issue: 02;
Start page: 25;
Date: 2013;
Original page

Keywords: 图像超分辨 | 图像超分辨应用 | 退化模型 | 学习模型 | 最大后验概率估计(MAP) | Image Super-Resolution | Application of Image Super-Resolution | Degradation Model | Learning Model | Maximum A Posteriori (MAP)

ABSTRACT
本文介绍图像超分辨的重建技术及其应用,图像超分辨技术的过程即获得比单幅图像更多的额外信息,然后通过特定的算法,把这些额外的信息融入到原来的图像当中,获得一幅高质量、高清晰度的图像。图像超分辨率重建技术可以分为基于重建的超分辨率重建技术和基于学习的超分辨率重建技术。基于重建的超分辨重建技术主要是依据退化模型,通过不同的算法来估计高分辨率图像。而基于学习的超分辨率重建技术则是从训练样本中获得先验信息,对原始低分辨率图像进行补充。本文详细介绍了基于重建和基于学习的超分辨重建技术的一些主要算法。随着人们对数字图像的分辨率要求越来越高,图像超分辨技术应用逐渐广泛,涉及军事、医学、银行、勘探等很多领域。Image super-resolution reconstruction technique and its application are introduced in this paper. Image super resolution technology process is to obtain more additional information than single image, and then with specific algorithm, the additional information is put into the original image to obtain an image with high quality and high definition. Super-resolution reconstruction includes two kinds of technology: reconstruction-based technology and learning-based technology. The reconstruction-based technology is estimating a high-resolution image from inputting images according to specific degradation model. Learning-based technology is supplying inputting images with prior knowledge from training examples. This paper introduces some major algorithms of reconstruction-based technology and learning-based technology in detail. As people are getting higher and higher requirement for the resolution of the digital image, image super resolution technology is applied more widely in many fields, such as military, medicine, bank, exploration, etc.

Tango Rapperswil
Tango Rapperswil

    
RPA Switzerland

RPA Switzerland

Robotic process automation