Abstract
In this paper, we present a novel onboard robust visual algorithm for long-term arbitrary 2D and 3D object tracking using a reliable global-local object model for unmanned aerial vehicle (UAV) applications, e.g., autonomous tracking and chasing a moving target. The first main approach in this novel algorithm is the use of a global matching and local tracking approach. In other words, the algorithm initially finds feature correspondences in a way that an improved binary descriptor is developed for global feature matching and an iterative Lucas–Kanade optical flow algorithm is employed for local feature tracking. The second main module is the use of an efficient local geometric filter (LGF), which handles outlier feature correspondences based on a new forward-backward pairwise dissimilarity measure, thereby maintaining pairwise geometric consistency. In the proposed LGF module, a hierarchical agglomerative clustering, i.e., bottom-up aggregation, is applied using an effective single-link method. The third proposed module is a heuristic local outlier factor (to the best of our knowledge, it is utilized for the first time to deal with outlier features in a visual tracking application), which further maximizes the representation of the target object in which we formulate outlier feature detection as a binary classification problem with the output features of the LGF module. Extensive UAV flight experiments show that the proposed visual tracker achieves real-time frame rates of more than thirty-five frames per second on an i7 processor with 640 × 512 image resolution and outperforms the most popular state-of-the-art trackers favorably in terms of robustness, efficiency and accuracy.
Highlights
Visual tracking, as one of the most active vision-based research topics, can assist unmanned aerial vehicles (UAVs) to achieve autonomous flights in different types of civilian applications, e.g., infrastructure inspection [1], person following [2] and aircraft avoidance [3]
Numerous visual tracking algorithms have recently been proposed in the computer vision community [4,5,6,7,8,9], onboard visual tracking of freewill arbitrary 2D or 3D objects for UAVs remains as a challenging task due to object appearance changes caused by a number of situations, inter alia, shape deformation, occlusion, various surrounding illumination, in-plane or out-of-plane rotation, large pose variation, onboard mechanical vibration, wind disturbance and aggressive UAV flight
A dendrogram generated from the single-link hierarchical agglomerative clustering (HAC) approach is shown in the middle of Figure 2; the Ck∪ is divided into some subgroups based on a cut-off threshold η, and the biggest subgroup is considered as the correspondences for the target object
Summary
As one of the most active vision-based research topics, can assist unmanned aerial vehicles (UAVs) to achieve autonomous flights in different types of civilian applications, e.g., infrastructure inspection [1], person following [2] and aircraft avoidance [3]. The following three main modules are proposed to have a reliable visual object model under the aforementioned challenging situations and to achieve those basic requirements in different UAV tracking applications:. An efficient local geometric filter (LGF) module has been designed for the proposed visual feature-based tracker to detect outliers from global and local feature correspondences, i.e., a novel forward-backward pairwise dissimilarity measure has been developed and utilized in a hierarchical agglomerative clustering (HAC) approach [14] to exclude outliers using an effective single-link approach. A heuristic local outlier factor (LOF) [15] module has been implemented for the first time to further remove outliers, thereby representing the target object in vision-based UAV tracking applications reliably.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
