Motion parameter estimation from global flow field data

Hummel, Robert A.; Sundareswaran, V.

doi:10.1109/34.211466

Cited by 34 publications

(15 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, I will focus on the problem of determining the observer's direction of motion. This problem has been studied extensively ever since Gibson's seminal work (Gibson 1950), and at present there are about a dozen different algorithms for computing heading [for a review see Warren et al (1988); more recent algorithms are given in Heeger and Jepson (1990), Burger and Bhanu (1990), Perrone (1992), and Hummel and Sundareswaran (1993)]. Although they differ in many respects, these algorithms all use cartesian ima#e flow, which measures the displacement of image features over time.…”

Section: Introductionmentioning

confidence: 98%

Computing the direction of heading from affine image flow

Beusmans

1993

Biol. Cybern.

View full text Add to dashboard Cite

Observers moving through a three-dimensional environment can use optic flow to determine their direction of heading. Existing heading algorithms use cartesian flow fields in which image flow is the displacement of image features over time. I explore a heading algorithm that uses affine flow instead. The affine flow at an image feature is its displacement modulo an affine transformation defined by its neighborhood. Modeling the observer's instantaneous motion by a translation and a rotation about an axis through its eye, affine flow is tangent to the translational field lines on the observer's viewing sphere. These field lines form a radial flow field whose center is the direction of heading. The affine flow heading algorithm has characteristics that can be used to determine whether the human visual system relies on it. The algorithm is immune to observer rotation and arbitrary affine transformations of its input images; its accuracy improves with increasing variation in environmental depth; and it cannot recover heading in an environment consisting of a single plane because affine flow vanishes in this case. Translational field lines can also be approximated through differential cartesian motion. I compare the performance of heading algorithms based on affine flow, differential cartesian flow, and least-squares search.

show abstract

Section: Introductionmentioning

confidence: 98%

Computing the direction of heading from affine image flow

Beusmans

1993

Biol. Cybern.

View full text Add to dashboard Cite

show abstract

“…The target/object is tracked by tracking its these kernels in consecutive image frames. This process is illustrated in (18).…”

Section: The Measurement Processmentioning

confidence: 99%

“…Comparing with existing active contour based method (3) (5) (15) , we only use the contour to initialize feature points and detect the corresponding converged points around the feature points instead of searching contour shape itself. Comparing with normal optical flow method (13) (18) , this method analyzes correspondence of feature points between frames in image sequence to decide feature points of a target. The interframe time interval also maybe not short enough.…”

Section: Introductionmentioning

confidence: 99%

Feature Point-Based Object Tracking

Wang

Sugisaka²,

Wang

2004

IEEJ Trans.EIS

View full text Add to dashboard Cite

In this paper, we describe a new algorithm for tracking a known static target in image frames from a moving platform. We have noticed that a image point is environment sensitive, but those changes of grouped points energy have their own statistical similarities in two image frames within limited time interval. This approach analyzes correspondence of interest points around every feature points between inter-frames in image sequence in order to decide those feature points. We tackle these tasks with three broad approaches. First, we make an active contour model of a target in order to build some low-energy feature points. The feature points give constraints of the input state space for interest point detection in an input image frame. The second step is a method of detecting interest points around every kernel. We take into account auto-correlation method to indicate the presence of interest points for the purpose of features state space in consecutive image frames that can be tracked. The third step of our strategy is to detect the correspondence of interest points by a probabilistic relaxation method in tracking windows. The detecting process is iterative and begins with the detection of all potential correspondence pair in consecutive image. Each pair of corresponding points is then iteratively recomputed to get a globally optimum set of pairwise correspondences. Successful results are given for a real vide frames.

show abstract

“…Due to its importance, many algorithms dealing with the problem of estimating egomotion have appeared in the literature. The following paragraphs provide a short review of a few representative methods; more detailed discussions can be found in [7,8,10]. Most of the methods reviewed here rely on the availability of a dense optical flow field to describe 2D motion.…”

Section: Introductionmentioning

confidence: 99%

“…Hummel and Sundareswaran [8] present an algorithm for finding the rotational motion and one for locating the FOE. The first algorithm is based on the observation that the curl of the optical flow field is approximately a linear function whose coefficients are proportional to the desired rotational parameters of motion.…”

Section: Introductionmentioning

confidence: 99%

Egomotion Estimation Using Quadruples of Collinear Image Points

Lourakis

2000

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. This paper considers a fundamental problem in visual motion perception, namely the problem of egomotion estimation based on visual input. Many of the existing techniques for solving this problem rely on restrictive assumptions regarding the observer's motion or even the scene structure. Moreover, they often resort to searching the high dimensional space of possible solutions, a strategy which might be inefficient in terms of computational complexity and exhibit convergence problems if the search is initiated far away from the correct solution. In this work, a novel linear constraint that involves quantities that depend on the egomotion parameters is developed. The constraint is defined in terms of the optical flow vectors pertaining to four collinear image points and is applicable regardless of the egomotion or the scene structure. In addition, it is exact in the sense that no approximations are made for deriving it. Combined with robust linear regression techniques, the constraint enables the recovery of the FOE, thereby decoupling the 3D motion parameters. Extensive simulations as well as experiments with real optical flow fields provide evidence regarding the performance of the proposed method under varying noise levels and camera motions.

show abstract

Motion parameter estimation from global flow field data

Cited by 34 publications

References 35 publications

Computing the direction of heading from affine image flow

Computing the direction of heading from affine image flow

Feature Point-Based Object Tracking

Egomotion Estimation Using Quadruples of Collinear Image Points

Contact Info

Product

Resources

About