In this work we present a robust detection method in outdoor scenes under cast shadows using color based invariant gradients in combination with HoG local features. The method achieves good detection rates in urban scene classification and person detection outperforming traditional methods based on intensity gradient detectors which are sensible to illumination variations but not to cast shadows. The method uses color based invariant gradients that emphasize material changes and extract relevant and invariant features for detection while neglecting shadow contours. This method allows to train and detect objects and scenes independently of scene illumination, cast and self shadows. Moreover, it allows to do training in one shot, that is, when the robot visits the scene for the first time.

This work was supported by projects: 'Ubiquitous networking robotics in urban settings' (E-00938), 'CONSOLIDER-INGENIO 2010 Multimodal interaction in pattern recognition and computer vision' (V-00069), 'Robotica ubicua para entornos urbanos' (J-01225), 'Percepción y acción ante incertidumbre' (4803). This research was partially supported by the Spanish Minist ry of Innovation and Science under projects Consolider Ingenio 2010 CSD2007-00018, and projects DPI 2007-614452 and DPI 2008-06022; by t he URUS project IST-045062 of the European Union; by the Generalita t of Catalonia’s Department of Innovation, University and Industry and the European Social Fund to JS; and by the Technical University of Catalonia (UPC) to MV.

Trabajo presentado al ICRA 2009 celebrado en Kobe (Japón) del 12 al 17 de mayo.

Peer Reviewed