Robust blind deconvolution process for vehicle re-identification by an Inductive Loop Detector

David Guilbert (1, 2) , Sio Song Ieng (3) , Cédric Le Bastard (1, 2) , Yide Wang (1)

1 IETR - Institut d'Électronique et des Technologies du numéRique
2 ENDSUM - Evaluation non destructive des structures et des matériaux
3 IFSTTAR/LEPSIS - Laboratoire Exploitation, Perception, Simulateurs et Simulations

A robust blind deconvolution algorithm is proposed to cancel the sensor averaging effect caused by its wide detection area. The purpose herein is to retrieve features of the "real" signal that have been distorted by the averaging effect. The algorithm is applied to the case of an Inductive Loop Detector. To perform the proposed algorithm, speed estimation is required. Vehicle reidentification rate from both raw signals and estimated "real" signals is compared. The sensor transfer function is calculated once from a learning phase; the estimated "real" signal is then computed in real time for the re-identification of each vehicle.

Domaines: Sciences de l'ingénieur [physics]; Traitement du signal et de l'image [eess.SP]

David Guilbert, Sio Song Ieng, Cédric Le Bastard, Yide Wang. Robust blind deconvolution process for vehicle re-identification by an Inductive Loop Detector. IEEE Sensors Journal, 2014, 14 (12), pp.4315-4322. ⟨10.1109/JSEN.2014.2345755⟩. ⟨hal-01058308⟩