Machine Vision Based Position Measurement for Autonomous Low Speed Flight

A real-time machine vision system is presented which enables an autonomous position measurement. This is done by combining a state estimation (4D-approach) with procedures enabling the detection and tracking of a priori unknown features. Thus, the system does not require additional knowledge about the appearance of the scenery seen by the camera. Besides the methods involved, this paper presents the hardware of the machine vision system as well as test results. The method can be used for applications requiring a relative position measurement like a helicopter hovering above a life boat, but the methodology is independent from the flight vehicle being used.