Video strip mosaicking: A two-dimensional approach by convergent image bridging

Abstract

A principal advantage of video is to provide, in a cost-effective manner, information required for a very narrow and long strip target, utilizing narrow view angle and dynamic stereo coverage. A video mosaicking technique has been specifically developed with a focus on such a narrow and long strip target (e.g. 20-30m wide and several hundred kilometres long). This technique was developed by combining traditional analogue mosaicking and digital image processing techniques. An interactive approach, registering video frames in two convergent directions along a strip, produced an acceptable positional accuracy in terms of time and cost requirements. The results also suggest that video strip mosaicking could effectively be used for linear targets which require generalized thematic mapping of relatively simple classes. This development of a simplified two-dimensional image bridging approach has enabled airborne videography to progress beyond being used mainly as a 'snapshot' visual assessment tool. The low-cost operational use of convergent image bridging, requiring solely endlap, identifies video strip mosaicking as a potentially useful remote sensing tool.