METHOD AND DEVICE FOR OBJECT RECOGNITION BY ANALYSIS OF DIGITAL IMAGE SIGNALS REPRESENTATIVE OF A SCENE

摘要

The invention relates to a method and a system for recognizing objects by analyzing digital image signals of a scene, said digital image signals being acquired by an image acquisition device mounted on a support comprising minus one axis of rotation. The object recognition method comprises steps of:acquiring (64) a first first resolution digital image signal corresponding to a wide field of view,calculating (66) a saliency map from the first digital image signal, comprising a prior visual saliency learning step (62) of a set of classes of objects of predetermined interest, at least an object image of each class being stored in a database,determining (68) at least one zone of interest from the calculated saliency map,-orientation (70) of the image acquisition device and acquisition of a second digital image signal focused on said second resolution area of interest greater than the first resolution,determining (72, 74, 76) an object classification of a part of the scene using the second digital image signal.