摘要
信息的不断多元化发展,使得图像已成为获取信息的重要手段之一,广泛应用于光学、摄影摄像、生产生活、军事及航空航天等领域。图像的质效决定信息获取的准确性,但是在实际生活中,获取的图像往往存在着一定程度的模糊。模糊的成因很多,例如大气模糊、运动模糊和噪声模糊等。由于模糊形成原因的多样性以及不确定性,使得对图像进行恢复非常复杂和困难,因此模糊图像恢复技术依然是当今的一个重要研究方向和技术难题。结合“针对自然图像的高速摄像机”项目,针对运动造成的模糊图像进行研究。目前对运动模糊图像的研究多是针对单幅图像匀速直线运动模糊,采用频域、时域、空域等办法估算其运动模糊参数,从而进行图像恢复,但是由于算法存在的误差以及图像运动模糊状态的不确定性,使得对模糊图像恢复的质量难以得到保证。为寻求最佳运动模糊恢复图像效果,研究将人机交互方式引入图像恢复中的可行性以及方法,结合现有针对单幅匀速直线运动模糊图像的恢复技术,设计实现运动模糊图像恢复系统,并针对恢复过程中出现的振铃效应进行消除。
As information has diversified into various forms,images have become an important means to retrieve information and are widely used in different domains such as optics,photography,life&production,military and aerospace applications.While the image quality often deter⁃mines the accuracy of information retrieval,the images taken in real life often exhibit certain degrees of ambiguity.There are various rea⁃sons that cause the ambiguity in images,such as atmospheric blur,motion blur and noise blur.Because of the multiplicity and uncertainty of the causes of image blur,it is complicated and challenging to recover blurred images.As such,fuzzy image restoration is still a technical challenge and important research topic.This article focuses on the study of motion-blurred images from high-speed video cameras.The ex⁃tant research on motion-blurred images has mostly focused on restoring individual blurred images caused by uniform linear motion by lever⁃aging frequency domain,time domain,and space domain approaches for motion blur parameter estimation.However,the quality of this kind of image restoration approach can be suboptimal because of the bias of the algorithm and the uncertainty of the degree of image motion blur.This article proposes a new image restoration approach that involves user interactions and investigates the feasibility of the approach.The proposed approach is able to restore motion-blurred images by incorporating existing restoration techniques for individual blurred im⁃ages caused by uniform linear motion as well as canceling the ring effect in the restoration process.
作者
吴垠
黄新
陈鹏
WU Yin;HUANG xin;CHEN Peng(General Office of Chengdu Municipal People's Government,Chengdu 610041;32620 Troops,PLA,Chengdu 610041;School of Computer and Software Engineering,Xihua University,Chengdu 610039)
出处
《现代计算机》
2020年第26期52-59,共8页
Modern Computer
关键词
图像模糊恢复
人机交互
用户体验
Image Fuzzy Restoration
Human Computer Interaction
User Experience
