UAV Attitude estimation by vanishing points in catadioptric images

Abstract

Abstract-Unmanned Aerial Vehicles (UAV) are the subject of an increasing interest in many applications and a key requirement is the stabilization of the vehicle. Some previous works have suggested using catadioptric vision, instead of traditional perspective cameras, in order to gather much more information from the environment and therefore improve the robustness of the UAV attitude estimation. This paper belongs to a series of recent publications of our research group concerning catadioptric vision for UAVs. Currently, we focus on the estimation of the complete attitude of a UAV flying in urban environment. In order to avoid the limitations of horizon-based approaches, the difficulties of traditional epipolar methods (such as rotation-translation ambiguity, lack of features, retrieving motion parameters from matrix decomposition, etc...) and improve UAV dynamic control, we suggest computing infinite homography. We show how catadioptric vision plays a key role to: first, extract a large number of lines, second robustly estimate the associated vanishing points and third, track them even during long video sequences. Therefore it is not only possible to estimate the relative rotation between consecutive frames but also compute the absolute rotation between two distant frames without error accumulation. Finally, we present some experimental results with ground truth data to demonstrate the accuracy and the robustness of our method.

A Generic Method of Line Matching for Central Imaging Systems under Short-Baseline Motion

“…For extracting line segments, we use the line extraction algorithm of [6]. For recovering R, along with our proposed method we also explain attitude estimation method of [10]. We will shortly give a brief explanation of these algorithms.…”

“…Therefore for estimating R, it is sufficient to have two enough distinct VP correspondences in the two views. Reference [10] has exploited above facts to estimate the rotation of an imaging system in two steps, extraction of vanishing points followed by recovering R by matching these points. For the sake of completeness, we briefly explain their method for extracting VP.…”

“…The geometric relation between two images required a priori here is the rotation of imaging system. This can directly be recorded during image acquisition or later by different available methods such as methods based on matching corresponding vanishing points [9,10]. In this work, we are interested in the last step of the matching algorithm.…”

Visual-inertial UAV attitude estimation using urban scene regularities

“…Alternatively, when the horizon is occluded and not visible, the vanishing points present in an urban scene have been analyzed to infer this invisible horizon [7] [8] [9]. For example, a set of mutually orthogonal vanishing points was found from parallel line bundles and sky detection identified the vertical vanishing point [8].…”

Particle Filter Approach Adapted to Catadioptric Images for Target Tracking Application

“…That is why more and more robots are equipped with catadioptric camera that can see in "almost every direction". It has been shown that these cameras can sensibly improve robustness and accuracy in robotic applications [3] [10]. Whereas there exists a large literature on target tracking by PF on traditional images, only a very few systems have been developed for catadioptric images.…”