Temporal synchronization of non-overlapping videos using known object motion

Abstract

a b s t r a c tThis paper presents a robust technique for temporally aligning multiple video sequences that have no spatial overlap between their fields of view. It is assumed that (i) a moving target with known trajectory is viewed by all cameras at non-overlapping periods in time, (ii) the target trajectory is estimated with a limited error at a constant sampling rate, and (iii) the sequences are recorded by stationary cameras with constant frame rates and fixed intrinsic and extrinsic parameters. The proposed approach reduces the problem of synchronizing N non-overlapping sequences to the problem of robustly estimating a single line from a set of appropriately-generated points in R Nþ1 . This line describes all temporal relations between the N sequences and the moving target. Our technique can handle arbitrarily-large misalignments between the sequences and does not require any a priori information about their temporal relations. Experimental results with real-world and synthetic sequences demonstrate that our method can accurately align the videos.

Phase-based spatio-temporal interpolation for accurate 3D localization in camera networks

“…Despite its higher precision, this latter method relies on image-based features extracted from the scene which, differently from ad-hoc devices, might not be always available and are subject to fluctuating levels of accuracy. A rather different approach is proposed by [6], where the knowledge of the trajectory of an object encompassing the view frustum of several cameras is exploited in order to synchronize them. A similar idea is also adopted in [5] where, in a complementary fashion, several mobile sensors are synchronized using simultaneous observation from the same camera.…”

Video motion stitching using trajectory and position similarities

“…Direct alignment uses all pixels in a video frame for synchronization and is suitable for videos containing changes in light; e.g., videos of fireworks. Featurebased alignment depends on features, such as points on moving objects or object trajectories, as a basis for the synchronization algorithm [17][18][19]. Wedge et al [17] presented a coarse-level to fine-level approach to synchronize two video sequences recorded at the same frame rate by stationary cameras with fixed internal parameters.…”

Video Synchronization Based on Projective-Invariant Descriptor

“…We address this problem by exploiting a novel projective-invariant descriptor based on the cross ratio to obtain the matched trajectory points between the two input videos. So far, numerous video synchronization methods have been presented in the previous works, which are mainly classified into two categories: intensity-based ones [9][10][11][12][13] and feature-based ones [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29]. The intensity-based methods usually rely on colors, intensities, or intensity gradients to achieve the temporal synchronization of overlapping videos.…”

“…Among the feature-based video synchronization methods, the trajectory-based ones are one of the most popular categories [19][20][21][22][23][24][25][26][27][28][29]. These methods generally use some epipolar geometry or homography information among different viewpoints for the purpose of exploiting the matched trajectory points or time pairs between (or among) input videos [21,22,26,28].…”