Innova Trajectory
THE DATABASE
PUBLICATIONS
VISIT ALSO
Innova_trajectory is a set of omnidirectional images captured in an indoor environment (ground floor, Innova Building) at Miguel Hernandez University, Spain. This database is intended to test visual mapping and localization algoritmhs for mobile robots in a real-working environment.
The images have been captured using a catadiptric vision system formed by a CCD color camera DFK-41BF02, pointing towards a hyperbolic mirror model Eizoh wide 70. The optical axis of the camera and the axis of the mirror are aligned. With this system, omnidirectional images with a resolution of 1280x960 pixels are captured. The length of the route is about 180m and the number of scenes in the set is 1450, with an average distance between capture positions around 12 cm. A Pioneer P3-AT robot was used to collect the set of images. It is equipped with a laser rangefinder and motor encoders, besides the catadioptric vision system. The laser data have been utilized to obtain the ground truth data (sing the software gmapping), which are also available.
The next figure shows a bird eye s view of the positions where the images were captured and some sample panoramic images captured in different rooms. Specifically, the robot navigates through a laboratory, some corredors and halls of the building. It is a challenging trajectory which was captured under real working conditions. There are many large windows that introduce lighting issues (shadows and reflections). Also, there are people normally moving along the environment while the robot captures the images. Besides, part of the path includes a slope so the information is captured in different projection planes. The slope is taversed both upwards and downwards. Finally, several loop closures occur.
The database includes the whole set of omnidirectional images and a file which includes the ground truth.
|
The whole database is downloadable through the next link: Innova Trajectory
Publications
Visit also
- ARVC (Automation, Robotics and Computer Vision group): Visit our research group.