Vision-Based Navigation for Autonomous Ground Vehicles
Abstract
This is a summary report for contract DACA 76-84-C-0004, Vision Based Navigation for Autonomous Ground Vehicles. Our research has resulted in seventeen technical reports (list appended to this report, with abstracts), many of which have been subsequently published in journals, conferences and workshops. Additionally, our project involved close collaboration with the Martin Marietta Corporation, Denver, Colorado, in the development and testing the vision algorithms for navigation of roads and road networks. Several experiments were run on the Martin Marietta Autonomous Land Vehicle using programs developed at the University of Maryland, and some critical components of Martin Marietta's visual navigation system were based on fundamental research conducted at the University of Maryland and under support of this contract specifically, the overall framework of a focus-of-attention vision system, in which detailed analyses are performed on selected windows of images of roads, and the shape-from-contour algorithms (e.g., the zero bank algorithm) that allowed the vehicle software to recover an accurate three-dimensional road model from monocular imagery, thus saving the autonomous land vehicle (ALV) from having to perform costly, and less reliable, analyses based on either stereo or motion.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1989
- Accession Number
- ADA211584
Entities
People
- Larry S. Davis
Organizations
- University of Maryland