ROBUST OBSTACLE DETECTION AND TRACKING BY MOTION ANALYSIS

In this paper, the authors describe an obstacle detection and tracking approach which evaluates motion information that is comprised in optical flow vectors. They discuss how optical flow contains information about camera motion relative to its environment and about the 3D structure of the imaged scene. Clusters of optical flow vectors are grouped, and a state estimation for each moving object is obtained using Kalman filtering. Implementation and experimental results from image sequences show the robustness of this obstacle detection approach.

Language

  • English

Media Info

  • Pagination: p. 717-722

Subject/Index Terms

Filing Info

  • Accession Number: 00801215
  • Record Type: Publication
  • Source Agency: UC Berkeley Transportation Library
  • Files: PATH
  • Created Date: Nov 7 2000 12:00AM