Skip to main content

Stereo Scene Flow for 3D Motion Analysis

  • Wedel A
  • Cremers D
Citations of this article
Mendeley users who have this article in their library.
Get full text


Most hazards in traffic situations include other moving traffic partici-pants. Hence, reliably detecting motion in world coordinates is a crucial step for many driver assistance systems and an important part of machine visual kines-thesia. In this chapter we present two extensions of scene flow, making it further amenable to practical challenges of driver assistance. Firstly, we present a frame-work for scene-flow-based moving object detection and segmentation. Secondly, we discuss the application of Kalman filters for propagating scene flow estimation over time. Without geometry life is pointless. Anonymous 6.1 Flow Cut—Moving Object Segmentation Classically, moving objects are separated from the stationary background by change detection (e.g. [89]). But if the camera is moving in a dynamic scene, motion fields become rather complex. Thus, the classic change detection approach is not suitable as can be seen in Fig. 6.1 (e.g. second from left). The goal in this chapter is to de-rive a segmentation of moving objects for the more general dynamic setting. The A. Wedel, D. Cremers, Stereo Scene Flow for 3D Motion Analysis,




Wedel, A., & Cremers, D. (2011). Stereo Scene Flow for 3D Motion Analysis. Stereo Scene Flow for 3D Motion Analysis. Springer London.

Register to see more suggestions

Mendeley helps you to discover research relevant for your work.

Already have an account?

Save time finding and organizing research with Mendeley

Sign up for free