US 11,054,946 B2
Image recognition device and touch panel
Pengpeng Wang, Beijing (CN); Haisheng Wang, Beijing (CN); Xiaoliang Ding, Beijing (CN); Chihjen Cheng, Beijing (CN); Yanling Han, Beijing (CN); and Ping Zhang, Beijing (CN)
Assigned to BOE TECHNOLOGY GROUP CO., LTD., Beijing (CN)
Filed by BOE TECHNOLOGY GROUP CO., LTD., Beijing (CN)
Filed on Dec. 28, 2018, as Appl. No. 16/234,953.
Claims priority of application No. 201820265645.1 (CN), filed on Feb. 23, 2018.
Prior Publication US 2019/0265839 A1, Aug. 29, 2019
Int. Cl. G06F 3/042 (2006.01); G06T 7/593 (2017.01); H04N 13/282 (2018.01); H04N 13/239 (2018.01); H04N 13/00 (2018.01)
CPC G06F 3/0421 (2013.01) [G06T 7/593 (2017.01); H04N 13/239 (2018.05); H04N 13/282 (2018.05); G06T 2207/10028 (2013.01); H04N 2013/0081 (2013.01)] 18 Claims
OG exemplary drawing
 
1. An image recognition apparatus comprising:
a plurality of depth cameras including a plurality of depth camera units, the plurality of depth camera units being arranged in an array on a same substrate,
wherein the plurality of depth cameras comprises at least two depth cameras which are configured to acquire local images of different depth ranges and different local field spaces for mosaicing; and
an image recognition processor connected respectively to the plurality of depth camera units for performing depth image calculation on multiple ones of the local images, and mosaicing in a depth direction and a field direction,
wherein the image recognition processor comprises:
a field mosaic sub-processor for mosaicing a plurality of local longitudinal depth images in the field direction,
wherein a distance between two depth camera units of each depth camera of the at least two depth cameras is different, so that a scene which a respective depth camera of the at least two depth cameras is suitable to measure has a different depth of field,
wherein the image recognition processor comprises:
a calculation sub-processor connected respectively to the plurality of depth camera units for performing depth image calculation on multiple ones of the local images by using a binocular depth detection algorithm to obtain a plurality of local depth images.