Method for estimating variational scene flow based on three-dimensional flow field regularization

Abstract

The invention discloses a method for estimating variational scene flow based on three-dimensional flow field regularization. The method comprises the following steps: utilizing calibrated left and right video cameras to obtain left and right image sequences; performing regularization on a three-dimensional flow field and obtaining a scene flow data item; extending a two-dimensional optical flow smooth item to a three-dimensional space and obtaining a scene-flow-driven anisotropic scene flow smooth item; designing a diffusion tensor and performing intrinsic decomposition according to direction information, obtaining diffusion intensity in each direction, performing anisotropic smoothing, and obtaining a depth smooth item; combining the scene flow data item, the scene flow smooth item and the depth smooth item, and creating an energy functional; using a variational minimization method and obtaining a solution of an Euler equation corresponding to the energy functional; utilizing super-relaxation iteration to perform iterative solving on the Euler equation and obtaining the scene flow and depth information after optimization. The method has the advantages of high robustness and accurate scene flow.

Description

technical field [0001] The invention belongs to a scene flow estimation method, in particular to a variational scene flow estimation method based on scene flow driving anisotropic smoothing and anisotropic depth map smoothing, and based on 3-dimensional flow field regularization. Background technique [0002] The real world is 3-dimensional. Traditional digital video records the projection of 3-dimensional dynamic scenes on a 2-dimensional plane, which appears as a 2-dimensional image sequence, which loses depth information and is not conducive to scientific research applications. Scene flow is an extension of optical flow from a 2-dimensional plane to a 3-dimensional space, which can describe the movement of scene points in 3-dimensional space. Optical flow can be regarded as the projection of scene flow on the 2D image plane, which is an approximation to the real motion field. Compared with optical flow, scene flow is more in line with 3D real motion. [0003] In the scen...

AI Technical Summary

Problems solved by technology

It has also been proposed to expand the 2D smoothing constraint to 3D, but the isotropic scene flow driving method is still used. The problem of a significant increase in scene flow estimation errors in occlusion or motion discontinuity areas still exists, and the anisotropic smoothing assumption It is important to overcome such problems

Images