Combination of Wide Baseline Image Matching and Tracking for Autonomous UAV Approaches to a Window

In this paper, the problem of an autonomous unmanned aerial vehicle (UAV) approach to a target which is selected in a reference image is addressed. A robust matching algorithm is proposed to reliably project the selected point in the reference image into the live images of a quadrotor helicopter. Projective transformations are applied to the reference image to extract additional keypoints and to gain invariance to out-of-plane perspective transformations. Since the matching algorithm does not exploit the beneficial characteristics of image sequences and because its processing time is not short enough for high frame rates, a tracking algorithm is introduced. High detection rates even for image sequences with large viewpoint changes are achieved. Therefore, the presented algorithm can be used as input to a guidance algorithm for UAVs.