Adaptation of heading angle for continuous visibility of tracks in geographical track following applications

Vision-based autonomous following of geographical tracks such as rivers and pipelines is a potential solution in applications denied by GPS. Vision-based track-following algorithms require two parameters to determine the aircraft heading namely, orientation of the track and the displacement of the track from an appropriate reference. Track visibility is a prime factor for effective mission. Despite clear visibility, tracks limited to corners of the image tend to be lost in future images after heading change. In this paper, we propose a post-processing technique to ensure continuous visibility of tracks. Experiments reveal an improvement in track followed by the existing algorithms.