Datasets

Long-term SLAM Dataset

Long-term or life-long SLAM (Simultaneous Localization and Mapping) has been increasing the interest of the robotics community in the last years. Application fields for SLAM range from service, industrial, security, inspection to space sectors. However, the aforementioned application fields also require an autonomous system deployed for long-term operation without human intervention. Therefore, testing new SLAM algorithms constrained to these requirements takes much time and effort. The collected dataset includes 15 runs of each floor of Building PIV at the University of Girona, covering a total distance of 550 m, 445 m and 640 m of the ground, first and second floor respectively. The data acquisitions were performed at different day times, and different year seassons. Also, the dataset includes 15 additional runs including a tour over all Building PIV. The data logged includes laser scans from a URG-04LX Hokuyo LRF, omnidirectional images from a UI-2230SE-C camera attached to a RemoteRality optics and the P3DX mobile robot odometry readings.

Laser Range Finder and Image Based Vehicle Classification Dataset

Today it is very important to know the number and type of vehicles on roadways. This information is used to record vehicular traffic data, which is a fundamental for Intelligent Transport Systems applications. Automatic vehicle classification systems (AVC) techniques have been widely considered to identify vehicles as they pass on highways without using any type of electronic pins or chips installed on the vehicle. AVC systems are used in many applications such as automatic tolls, bridge/tunnel clearance verification and road surveillance. This dataset includes range images of different kind of vehicles captured at one of the Universidad del Valle entry points, the vehicle estimated velocity, and each range image was labeled with the corresponding class for validation purposes. The sensor used was a SICK LRF LMS200.

In the same way, the image based vehicle classification dataset was also acquired. This dataset includes different visual images, vehicle velocities and labeling information corresponding to each class. The images were acquired using a Canon VC-C50i camera, providing images of 720x450 pixels.