Combining two-view constraints for motion estimation

Govindu, Venu Madhav

doi:10.1109/cvpr.2001.990963

Cited by 166 publications

(225 citation statements)

References 12 publications

Supporting

Mentioning

225

Contrasting

Order By: Relevance

“…For low levels of noise we will use the method proposed in [6]. It is based on representing rotations as unit quaternions.…”

Section: Handling Low-level Noisementioning

confidence: 99%

“…In [6] the camera orientations are determined by solving these equations in a least squares sense. To motivate this, look at two unit quaternions p and q corresponding to rotations R p and R q such that d(R p , R q ) = α.…”

Section: Handling Low-level Noisementioning

confidence: 99%

“…Instead we will have to deal with low-level noise as well as completely inconsistent rotations. Low-level noise can be handled with the method in [6] but first, all outlier rotations must be removed. The This experiment shows that there is no reason not to use view pairs with short baselines.…”

Section: Problem Formulationmentioning

confidence: 99%

See 2 more Smart Citations

Non-sequential structure from motion

Enqvist

Kahl

Olsson

2011

2011 IEEE International Conference on Computer Vision Workshops (ICCV Workshops)

122

View full text Add to dashboard Cite

Prior work on multi-view structure from motion is dominated by sequential approaches starting from a single twoview reconstruction, then adding new images one by one. In contrast, we propose a non-sequential methodology based on rotational consistency and robust estimation using convex optimization. The resulting system is more robust with respect to (i) unreliable two-view estimations caused by short baselines, (ii) repetitive scenes with locally consistent structures that are not consistent with the global geometry and (iii) loop closing as errors are not propagated in a sequential manner. Both theoretical justifications and experimental comparisons are given to support these claims.

show abstract

“…For low levels of noise we will use the method proposed in [6]. It is based on representing rotations as unit quaternions.…”

Section: Handling Low-level Noisementioning

confidence: 99%

Section: Handling Low-level Noisementioning

confidence: 99%

See 1 more Smart Citation

Non-sequential structure from motion

Enqvist

Kahl

Olsson

2011

2011 IEEE International Conference on Computer Vision Workshops (ICCV Workshops)

122

View full text Add to dashboard Cite

show abstract

“…Early work on global image orientation has been conducted by (Govindu, 2001). In the majority of existing approaches absolute camera rotations are estimated and held fix for a subsequent separate estimation of camera poses.…”

Section: Related Workmentioning

confidence: 99%

Hierarchical Structure From Motion Combining Global Image Orientation and Structureless Bundle Adjustment

Cefalu

Haala

Fritsch

2017

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:Global image orientation techniques aim at estimating camera rotations and positions for a whole set of images simultaneously. One of the main arguments for these procedures is an improved robustness against drifting of camera stations in comparison to more classical sequential approaches. Usually, the process consists of computation of absolute rotations and, in a second step, absolute positions for the cameras. Either the first or both steps rely on the network of transformations arising from relative orientations between cameras. Therefore, the quality of the obtained absolute results is influenced by tensions in the network. These may e.g. be induced by insufficient knowledge of the intrinsic camera parameters. Another reason can be found in local weaknesses of image connectivity. We apply a hierarchical approach with intermediate bundle adjustment to reduce these effects. We adopt efficient global techniques which register image triplets based on fixed absolute camera rotations and scaled relative camera translations but do not involve scene structure elements in the fusion step. Our variant employs submodels of arbitrary size, orientation and scale, by computing relative rotations and scales between -and subsequently absolute rotations and scales for -submodels and is applied hierarchically. Furthermore we substitute classical bundle adjustment by a structureless approach based on epipolar geometry and augmented with a scale consistency constraint.

show abstract

“…The most common approach to rotation averaging are variants of [13], where a relaxation of the maximum likelihood estimator is solved. However, it has recently been shown [8] that this approach may fail if there is a camera loop where the total rotation is 360 degrees.…”

Section: Robust Global Rotation Computationmentioning

confidence: 99%

Stable Structure from Motion for Unordered Image Collections

Olsson

Enqvist

2011

Lecture Notes in Computer Science

103

View full text Add to dashboard Cite

Abstract. We present a non-incremental approach to structure from motion. Our solution is based on robustly computing global rotations from relative geometries and feeding these into the known-rotation framework to create an initial solution for bundle adjustment. To increase robustness we present a new method for constructing reliable point tracks from pairwise matches. We show that our method can be seen as maximizing the reliability of a point track if the quality of the weakest link in the track is used to evaluate reliability. To estimate the final geometry we alternate between bundle adjustment and a robust version of the known-rotation formulation. The ability to compute both structure and camera translations independent of initialization makes our algorithm insensitive to degenerate epipolar geometries. We demonstrate the performance of our system on a number of image collections. 1

show abstract

Combining two-view constraints for motion estimation

Abstract: In

Cited by 166 publications

References 12 publications

Non-sequential structure from motion

Non-sequential structure from motion

Hierarchical Structure From Motion Combining Global Image Orientation and Structureless Bundle Adjustment

Stable Structure from Motion for Unordered Image Collections

Contact Info

Product

Resources

About