A. M. Waxman and S. Sinha, Dynamic stereo: Passive ranging to moving objects from relative image flow, IEEE Trans. Pattern Anal. Mach. Intelligence, Vol. 8, 1989, pp. 406-412  Home/Search   Document Not in Database   Summary   Related Articles   Check

....scene. Moreover, no absolute solution of 3 D motion and structure can be obtained from a monocular image sequence by any above method. In order to avoid scalar ambiguities and improve the estimation performance, the dynamic stereo is thus introduced by the integration of stereo and SFM methods [Mit84, Ric85, WS86, WD86]. However, they need matching two monocular optical flows, which is very complex. In this paper we develop a new algorithm which integrates the methods of stereo and SFM, where monocular optical flows are not needed to match and the above difficulties of the integration are avoided. The 3 D ....

A. M. Waxman and S. Sinha, Dynamic stereo: Passive ranging to moving objects from relative image flow, IEEE Trans. Pattern Anal. Mach. Intelligence, Vol. 8, 1989, pp. 406-412

....rotation and rotational acceleration) over time [1, 2] A more complete set of parameters, which includes translational speed and acceleration and absolute depth, cannot be recovered without resorting to stereo imagery. There is a smaller body of work on the use of binocular image velocity fields [19, 27, 13, 3, 28, 14, 24, 12, 20, 21, 32, 11]. One approach uses known optical flow and point correspondence to recover 3D motion and depth [19, 27] difference in image velocity in the left and right images for the same image location) and disparity to compute 3D information. Another approach uses relative image velocity differences in the ....

....and absolute depth, cannot be recovered without resorting to stereo imagery. There is a smaller body of work on the use of binocular image velocity fields [19, 27, 13, 3, 28, 14, 24, 12, 20, 21, 32, 11] One approach uses known optical flow and point correspondence to recover 3D motion and depth [19, 27]. difference in image velocity in the left and right images for the same image location) and disparity to compute 3D information. Another approach uses relative image velocity differences in the left and right flow fields for the same image locations and disparity information to compute 3D ....

A. M. Waxman and S. S. Sinha. Dynamic stereo: Passive ranging to moving objects from relative image flows. IEEE PAMI, 8(4):406--412, 1984.

....established, and 3 D velocity vector fields are reconstructed. Mitiche [82] has investigated multiple optical flow based estimation. He computed optical flow for each image and via some assumptions was able to estimate depth and other motion parameters in a straightforward manner. Waxman and Sinha [143] have used a similar approach using filtering of the optic flow to reduce the effects of noise. They introduced a new concept in passive ranging, termed Dynamic Stereo. Several other researchers have also suggested new approaches [10] 3.4 On Recent Approaches Previous sections introduced the ....

Waxman, A.M., Sinha, S.S., "Dynamic Stereo: Passive Ranging to Moving Objects from Relative Image Flows", IEEE Transactions on Pattern Analysis and Machine Vision, vol. 8, no.4, 1986, pp. 406-412.

....of static stereopsis to temporal correspondence adds versatility to the solutions of several related problems. This method has been known in literature by various names like dynamic stereo, stereo motion analysis, binocular motion and so on. Significant studies of this process have been made by [1, 2, 3, 4]. In this paper we present the lower bound for errors in depth in the case of point or feature correspondence algorithms (under known motion and calibrated stereo camera systems) for binocular motion. As far as monocular motion is concerned far more studies have been done when compared to ....

A. M. Waxman and S. Sinha. Dynamic stereo: Passive ranging to moving objects from relative image flows. IEEE Transactions on Pattern Analysis and Machine Intelligence, (vol 8 no 4), 1986.

....that two pairs of homologous points for which optical flow is known are sufficient to recover the kinematic screw. The number of unknowns can be reduced when the type of motion is restricted to, for example, a simple translation [85, 100] For a system of parallel cameras, this constraint becomes [155, 157]: Deltau x = f 2 b Z r z ffi 2 x fb Z (y l r x Gamma x l r y ) 61) Here, even if the separation between the two cameras is not known, two pairs of homologous points are sufficient to obtain structure (up to a scale factor) 123] 6 Coding of stereokinetic images Large amounts ....

A.M. Waxman and S.S. Sarvajit. Dynamic stereo: Passive ranging to moving objects from relative image flows. IEEE Transactions on Pattern Analysis and Machine Intelligence, 8(4):406--412, july 1986.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC