FRIDA (Foggy Road Image DAtabase)

Image database for the evaluation of visibility restoration algorithms

FRIDA is a database of numerical images easily usable to evaluate in a systematic way the performance of visibility and contrast restoration algorithms. It comprises 90 synthetic images of diverse road scenes, taken in a view point close to the one of the vehicle’s driver, with and without fog. To each image without fog is associated 4 foggy images and a depthmap. Different kind of fog are added on each of the 4 associated images: uniform fog, heterogeneous fog, cloudy fog, and cloudy heterogeneous fog. These 18 scenes can be used to test visibility and contrast restoration algorithms intensively and in an objective way, as well as shape from fog algorithms. The calibration parameters of the camera are given. The image database can be downloaded as a unique archive file and used for research purpose:

(Size: 34Mo, Update: 2 juin 2010)

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

J.-P. Tarel, N. Hautière, A. Cord, D. Gruyer and H. Halmaoui, "Improved Visibility of Road Scene Images under Heterogeneous Fog", in Proceedings of IEEE Intelligent Vehicles Symposium (IV’10), San Diego, CA, USA, June 21-24, 2010. Bibtex Pdf

The images without fog are generated using Sivic Software.

The authors of this database are:

• Jean-Philippe Tarel (LEPSiS),
• Aurélien Cord (LIVIC),
• Houssam Halmaoui (LIVIC),
• Dominique Gruyer (LIVIC),
• Nicolas Hautière (LEPSiS).

Any query about the ownership of the FRIDA database should be addressed to Jean-Philippe.Tarel@lcpc.fr.