Accurate estimation of object location in an image sequence using helicopter flight data

✍ Scribed by Yuan-Liang Tang; Rangachar Kasturi

• Publisher: Elsevier Science
• Year: 1994
• Tongue: English
• Weight: 705 KB
• Volume: 11
• Category: Article
• ISSN: 0736-5845
• DOI: 10.1016/0736-5845(94)90010-8

✦ Synopsis

In automomous navigation, it is essential to obtain a three-dimensional (3D) description of the static environment in which the vehicle is traveling. For rotorcraft conducting low-altitude flight, this description is particularly useful for obstacle detection and avoidance . In this paper, we address the problem of 3D position estimation for static objects from a monocular sequence of images captured from a low-altitude flying helicopter . Since the environment is static, it is well known that the optical flow in the image will produce a radiating pattern from the focus of expansion . We propose a motion analysis system that utilizes the epipolar constraint to accurately estimate 3D positions of scene objects in a real world image sequence taken from a low-altitude flying helicopter . Results show that this approach gives good estimates of object positions near the rotorcraft's intended f ightpath .