Hakeem-Geospatial-617

Hakeem-Geospatial-617 - Estimating Geospatial Trajectory of...

Info iconThis preview shows pages 1–2. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Estimating Geospatial Trajectory of a Moving Camera Asaad Hakeem 1 , Roberto Vezzani 2 , Mubarak Shah 1 , Rita Cucchiara 2 1 School of Electrical Engineering and Computer Science, University of Central Florida, Orlando, Florida 32816, USA. { ahakeem,shah } @cs.ucf.edu 2 Dipartimento di Ingegneria dellInformazione, Universit di Modena e Reggio Emilia, Modena 41100, Italy. { vezzani.roberto,cucchiara.rita } @unimore.it Abstract This paper proposes a novel method for estimating the geospatial trajectory of a moving camera. The proposed method uses a set of reference images with known GPS (global positioning system) locations to recover the trajec- tory of a moving camera using geometric constraints. The proposed method has three main steps. First, scale invari- ant features transform (SIFT) are detected and matched be- tween the reference images and the video frames to calcu- late a weighted adjacency matrix (WAM) based on the num- ber of SIFT matches. Second, using the estimated WAM, the maximum matching reference image is selected for the cur- rent video frame, which is then used to estimate the relative position (rotation and translation) of the video frame using the fundamental matrix constraint. The relative position is recovered upto a scale factor and a triangulation among the video frame and two reference images is performed to resolve the scale ambiguity. Third, an outlier rejection and trajectory smoothing (using b-spline) post processing step is employed. This is because the estimated camera loca- tions may be noisy due to bad point correspondence or de- generate estimates of fundamental matrices. Results of re- covering camera trajectory are reported for real sequences. 1. Introduction GPS was first introduced by the US Department of De- fense (DoD) about 15 years ago for military personnel and vehicle tracking around the world. Since then the GPS tech- nology has been widely used in the areas of autonomous navigation and localization of vehicles and robots. Re- cently, commercial applications have employed GPS data with georeferenced maps to recover the map of a city ad- dress or to provide directions between different city lo- cations. In this paper, we address two issues: Localiz- ing the geospatial position and estimating the trajectory of a moving camera based on the captured sequence of im- ages. Geospatial position is localized using maximum SIFT matches between the reference images (with known GPS) and video frames while a camera trajectory is estimated us- ing the geometric constraint between maximum matching reference images and video frames. In literature, a variety of methods have been proposed for motion recovery and measurement of robot trajectory (odometry) using visual inputs. Structure from Motion (SFM) is the most common approach to solve problems such as automatic environment reconstruction, autonomous robot navigation and self-localization. These approaches employ a 3D reconstruction of the environment during learning phase or directly use the test video. Thus, the ac-learning phase or directly use the test video....
View Full Document

Page1 / 6

Hakeem-Geospatial-617 - Estimating Geospatial Trajectory of...

This preview shows document pages 1 - 2. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online