US 12,462,583 B2
Method for recognizing activated lamps at a vehicle
Fabian Zoepf, Geislingen an der Steige (DE)
Assigned to ROBERT BOSCH GMBH, Stuttgart (DE)
Filed by Robert Bosch GmbH, Stuttgart (DE)
Filed on Jan. 5, 2023, as Appl. No. 18/150,287.
Claims priority of application No. 10 2022 200 136.2 (DE), filed on Jan. 10, 2022.
Prior Publication US 2023/0222812 A1, Jul. 13, 2023
Int. Cl. G06V 20/58 (2022.01); G06V 10/22 (2022.01); G06V 10/764 (2022.01); G06V 10/776 (2022.01); G06V 10/82 (2022.01)
CPC G06V 20/584 (2022.01) [G06V 10/22 (2022.01); G06V 10/764 (2022.01); G06V 10/776 (2022.01); G06V 10/82 (2022.01)] 14 Claims
OG exemplary drawing
 
1. A method for recognizing which lamps at a vehicle are activated, including the steps:
providing multiple image recordings of candidate areas at the vehicle in which an activated lamp is presumed;
converting the image recordings into an intermediate product by executing a recurrent encoder network (ERNN) multiple times, the output of at least one pass of the ERNN being supplied as input to a further pass of the ERNN, and different image recordings of candidate areas being supplied as input to different passes of the ERNN; and
ascertaining assignments of the image recordings of candidate areas to classes which represent specific lamps of the vehicle from the intermediate product by executing a recurrent decoder network (DRNN) multiple times, the output of at least one pass of the DRNN is supplied as input to a further pass of the DRNN, and in each pass of the DRNN, an assignment of the image recording which was processed in the pass of the ERNN corresponding to the pass of the DRNN is ascertained to at least one class.