US 11,941,863 B2
Imaging system and method using a multi-layer model approach to provide robust object detection
Roberto Musiani, Monte San Pietro (IT); Alessandro Franchi, Bologna (IT); Maurizio De Girolami, Calderara di Reno (IT); and Angelo Carraggi, Zola Predosa (IT)
Assigned to Datalogic IP Tech S.r.l., Lippo de Calderara di Reno (IT)
Filed by Datalogic IP Tech S.r.l., Lippo de Calderara di Reno (IT)
Filed on Aug. 4, 2021, as Appl. No. 17/394,358.
Prior Publication US 2023/0038286 A1, Feb. 9, 2023
Int. Cl. G06V 10/75 (2022.01); G06F 18/22 (2023.01); G06V 10/44 (2022.01); G06V 10/74 (2022.01); G06V 30/24 (2022.01)
CPC G06V 10/751 (2022.01) [G06F 18/22 (2023.01); G06V 10/44 (2022.01); G06V 10/74 (2022.01); G06V 30/2504 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A computer-implemented method of detecting an image of a template object in a captured image, the method comprising:
comparing, by a processor, an image model of an imaged template object to multiple locations, rotations, and scales in the captured image, the image model defined by multiple model base point sets derived from contours of the imaged template object, each model base point set inclusive of a plurality of model base points that are positioned at corresponding locations associated with distinctive features of the imaged template object, and each corresponding model base point of the model base point sets (i) being associated with respective layers and (ii) having an associated gradient vector;
determining, by the processor, based on the comparison of the image model to the multiple locations, rotations, and scales in the captured image, whether and where the image of the object described by the image model is located in the captured image;
computing a multi-layer similarity metric by evaluating, for each model point, a score curve composed by all the scores of each point included in a corresponding base point set;
performing a peak detection of the score curve to identify a best matching layer of the model base point sets; and
filtering the score curve using a probabilistic filter to smoothen the score curve prior to performing the peak detection.