Download of the full VeLaSCa database

The database VeLaSCa simulates the journey of an egovehicle moving on a two-lane road with 5 vehicles used as targets, on the the Satory (Versailles, France) test track. It is a test road of 3.5 km with a 3D topography accurately
measured by professional land surveyors and used in order to build the Satory’s 3D realistic virtual environment. The VeLaSCa test is run during 165 seconds on a subpart of the Satory’s track.

The following data are logged:

• camera data with grey level 640 × 480 images at 25 img/s: a total of 4100 images are recorded, some examples are illustrated on Fig. 8 at different times.

• laser scanner data from the ego-vehicle: this sensor scans a 100o angle aperture with a 80 meters of range, a 0.25o resolution, and a 25Hz frequency. This sensor provides at each time a frame of 401 laser impacts. The detected and filtered objects coming from the laser data processing are displayed in Fig. 6 (red points).

• current positioning (X0, Y0, Z0) and heading of the egovehicle in the world frame.

• current positioning for each obstacle in both the world frame and the relative ego-vehicle referential. Datas (Xk ? X0, Yk ? Y0) for k = [1...5] are the record of the longitudinal (depth) and lateral relative position of the obstacles.

The images and datas are generated using Sivic Software.

The authors of this database are:

• Dominique Gruyer.
• Rachid Belaroussi

In any publication related to the use of the VeLaSCa image database, your are kindly requested to cite the following reference:

[1] Dominique Gruyer, Aurélien Cord and Rachid Belaroussi. Vehicle Detection and Tracking by Collaborative Fusion Between  Laser Scanner and  Camera. IROS 2013