Extracting Differential Invariants of Motion Directly From Optical Flow

The structure-from-motion problem is traditionally addressedbyestablishingpointcorrespondencesandthenapplying classic geometry techniques. However, depending on theapplication,perhapsonlypartialscenereconstructionis necessary. In such cases, computing the flrst-order difierential invariants of image motion, namely divergence, curl, and deformation, can directly provide information about scene structure, while avoiding complex projective geometry. Even though divergence, curl, and deformation have been shown to be useful for partial scene reconstruction, little work has been done to extract these quantities from imagesequences. Inthispaperweproposeawaytoextract the difierential invariants of image motion from an optical ∞owfleldusingabankoffllters. Theoutputofthesefllters can later be used for the recovery of surface normals and time-to-contact.

[1]  David J. Fleet,et al.  Performance of optical flow techniques , 1994, International Journal of Computer Vision.

[2]  Andrew Blake,et al.  Image Divergence and Deformation from Closed Curves , 1997, Int. J. Robotics Res..

[3]  Ruzena Bajcsy,et al.  Recent Advances in Motion Understanding , 1998, IVCNZ.

[4]  David Suter,et al.  Robust Optic Flow Computation , 1998, International Journal of Computer Vision.

[5]  David J. Fleet,et al.  On optical flow , 1995 .

[6]  Andrea J. van Doorn,et al.  Invariant Properties of the Motion Parallax Field due to the Movement of Rigid Bodies Relative to an Observer , 1975 .

[7]  Steven S. Beauchemin,et al.  The computation of optical flow , 1995, CSUR.

[8]  David J. Fleet.  Measurement of image velocity , 1992 .

[9]  Yiannis Aloimonos,et al.  Obstacle Avoidance Using Flow Field Divergence , 1989, IEEE Trans. Pattern Anal. Mach. Intell..

[10]  Changming Sun,et al.  Fast optical flow using 3D shortest path techniques , 2002, Image Vis. Comput..

[11]  Ta Camus,et al.  Real-time quantized optical flow , 1995, Proceedings of Conference on Computer Architectures for Machine Perception.

[12]  Nicola Ancona,et al.  Optical flow from 1-D correlation: Application to a simple time-to-crash detector , 1993, 1993 (4th) International Conference on Computer Vision.

[13]  Berthold K. P. Horn.  Robot vision , 1986, MIT electrical engineering and computer science series.

[14]  Mandyam V. Srinivasan,et al.  Measurement of optical flow by a generalized gradient scheme , 1991 .

[15]  Andrew Blake,et al.  Surface Orientation and Time to Contact from Image Divergence and Deformation , 1992, ECCV.

[16]  Roberto Cipolla,et al.  Active Visual Inference of Surface Shape , 1995, Lecture Notes in Computer Science.

[17]  Y. V. Venkatesh,et al.  Measurement of complex optical flow with use of an augmented generalized gradient scheme , 1994 .