Abstract : The proposed algorithm is a fast evolution-based vision technique for real-time obstacle detection. Based on the Parisian approach, our algorithm evolves a population of 3D particles which constitutes a three-dimensional representation of the scene. Evolution is controlled by a fuzzy fitness function able to deal with uncertain camera measurements, and uses classical evolutionary operators. The result of the algorithm is a set of 3D particles gathered on the surfaces of obstacles.
https://hal-lirmm.ccsd.cnrs.fr/lirmm-00954358 Contributor : Haythem GhazouaniConnect in order to contact the contributor Submitted on : Saturday, March 1, 2014 - 5:45:37 PM Last modification on : Friday, August 5, 2022 - 3:03:31 PM Long-term archiving on: : Thursday, May 29, 2014 - 2:30:16 PM
Haythem Ghazouani, Tagina Moncef, René Zapata. Evolution-Based Vision Algorithm with Fuzzy Fitness Function for Obstacle Detection. META: Metaheuristics and Nature Inspired Computing, Oct 2010, Djerba, Tunisia. pp.51-52. ⟨lirmm-00954358⟩