首页>
外国专利>
Method and system for real-time motion artifact handling and noise removal for runtime sensor images
Method and system for real-time motion artifact handling and noise removal for runtime sensor images
展开▼
机译:用于运行时传感器图像的实时运动伪影处理和噪声去除的方法和系统
展开▼
页面导航
摘要
著录项
相似文献
摘要
The invention relates to a method and system for real-time motion artifact handling and noise removal for runtime sensor (TOF) sensor images, wherein the TOF sensor images correspond to received signals (r (t)) received from a TOF sensor the transmission of modulated transmitted signals (s (t)). The method comprises: calculating values of a cross-correlation function C (T) at a plurality of temporally spaced positions or phases from the transmitted (s (t)) and received (r (t)) signals, whereby a plurality of respective cross-correlation values [C (T0 ), C (T1), C (T2), C (T3)]; Deriving from the plurality of cross-correlation values [C (T0), C (T1), C (T2), C (T3)], a depth image D, wherein the depth image D has values representing the distance to a part of a pixel for each pixel Representing object to which the transmitted signals (s (t)) are incident; Deriving from the plurality of cross-correlation values [C (T0), C (T1), C (T2), C (T3)], a guide image (I; I '), the guide image (I; I') comprising an image well defined edges; and generating an output image D 'on the basis of the depth image D and the guide image (I; I'), the output image D 'being an edge-preserving and smoothed version of the depth image D, the edge preservation being from the guide image (I; I'). The method may further include: filtering the guide image I to produce a noisy guide image t; and generating an output image D 'on the basis of the depth image D and the denuded guide image f. The method may further comprise: generating a plausibility image P based on the plurality of cross-correlation values [C (T0), C (T1), C (T2), C (T3)], wherein the plausibility image P for each pixel of the depth image D is a Includes value indicating whether the pixel is corrupted by motion or saturation; and generating the output image D 'on the basis of the depth image D and the plausibility image P and of either the guide image I or the noise image guide f.
展开▼