Optical Attitude Determination From Horizon Orientation Using Image Segmentation

摘要

Images from cameras onboard an aircraft or spacecraft frequently contain a wealth of information that may be exploited for navigation and other purposes. One feature that may be used for aircraft navigation in many situations is the orientation of the horizon. If the area around which the vehicle is flying is known to have little local topography, then the orientation of the observed horizon in a camera image may be used to estimate vehicle pitch and roll. This paper presents a new approach for detecting the horizon based on image segmentation along with a detailed discussion of the optimality of the segmentation result The technique developed here is a contributing algorithm to the miniature autonomous flight testbed.