Geolocalization with aerial image sequence for UAVs
- PDF / 8,864,328 Bytes
- 17 Pages / 595.276 x 790.866 pts Page_size
- 2 Downloads / 221 Views
Geolocalization with aerial image sequence for UAVs Yongfei Li1 · Hao He1 · Dongfang Yang1
· Shicheng Wang1 · Meng Zhang1
Received: 9 July 2019 / Accepted: 19 June 2020 © Springer Science+Business Media, LLC, part of Springer Nature 2020
Abstract The estimation of geolocation for aerial images is significant for tasks like map creating, or automatic navigation for unmanned aerial vehicles (UAVs). We propose a novel geolocalization method for the UAVs using only aerial images and reference road map. The corresponding road maps of the aerial images are firstly merged into a whole mosaic image using our newlydesigned aerial image mosaicking algorithm, where the relative homography transformations between road images are firstly estimated using keypoints tracking in RGB aerial images, and then further refined with registration between detected roads. The geolocalization of the aerial mosaic image is then taken as the problem of registering observed roads in the aerial images to the reference road map under the homography transformation. The registration problem is solved with our fast search algorithm based on a novel projective-invariant feature, which consists of two road intersections augmented with their tangents. Experiments demonstrate that the proposed method can localize the aerial image sequence over an area larger than 1000 km2 within a few seconds. Keywords Geolocalization · Road map mosaicking · 2-Road-intersections-tuple · GIS
1 Introduction It is of great importance for the automatic navigation of UAVs to estimate accurate global position. Nowadays, the most widely used global positioning methods are GPS based methods. However, there are some shortcomings in these methods, such as the multipath effect, making it unsuitable for some scenes. With the increasing use of cameras, localizing globally with cameras has attracted more and more attention. Generally speaking, vision-based geolocalization can be divided into two categories, according to the reference database used. One is the aerial images based method (Shan et al. 2015; Conte and Doherty 2008), and the other is the
B
Dongfang Yang [email protected] Yongfei Li [email protected] Hao He [email protected] Shicheng Wang [email protected] Meng Zhang [email protected]
1
Department of Control Engineering, Xi’an Hi-tech Research Institute, Xi’an City, Shaanxi Province, China
Geographic Information System (GIS) data based method. Many researches focus on the first kind of methods. Among them, bag-of-words (BoWs) is the most commonly used framework (Divecha and Newsam 2016). In these methods, reference aerial images are first converted into visual word descriptors to form a search tree. And then query images are also described with visual word descriptors. After that, all the best image candidates in the reference database are picked by retrieving in the search tree. At last, a geometric verification is performed to find the best matching image. Although effective, such methods are still faced with many challenges. First, to obtain the precise locat
Data Loading...
