Blind Deconvolution

Blind Deconvolution
复制标题

DOI:
10.1007/978-3-030-03243-2_771-1
复制
发表时间:
2021
期刊:
Computer Vision
影响因子:
--
通讯作者:
Tom E. Bishop;P. Favaro
Tom E. Bishop;P. Favaro
中科院分区:
其他
文献类型:
--
作者:
Tom E. Bishop;P. Favaro

文献摘要

被引文献

相似文献

背景正确的摄影曝光需要在曝光时间和光圈设置之间进行权衡。当照度较差时,摄影师可以选择使用长曝光时间或大光圈。当摄影机在曝光过程中相对于场景中的对象移动时,第一个设置会导致运动模糊。第二个设置会导致远离焦平面的深度处的对象出现焦外模糊。此外,由于相机抖动、不正确的对焦设置或诸如大气湍流之类的其他失真,用户可能会加剧这些效果。在局部近似下,这些过程可以被建模为理想的“锐化”图像和核或点扩展函数(PSF)之间的卷积运算。该PSF表示场景中的一个光点将如何成像到相机的传感器上。通常,PSF可以是空间变化的或深度变化的,使得3D空间中的每个点在
BackgroundA correct photographic exposure requires a tradeoff in exposure time and aperture setting. When illumination is poor, the photographer can choose to use a long exposure time or a large aperture. The first setting results in motion blur when the camera moves relative to objects in the scene during the exposure. The second setting results in out-of-focus blur for objects at depths away from the focal plane. Furthermore, these effects may be exacerbated by the user due to camera shake, incorrect focus settings, or other distortions such as atmospheric turbulence. Under local approximations, these processes can be modeled as convolution operations between an ideal “sharp” image and a kernel or point spread function (PSF). This PSF represents how a point of light in the scene would be imaged onto the camera’s sensor. In general, the PSF may be space-varying or depth-varying, such that each point in 3D space has its own response in