US 12,230,032 B2
Item identification and tracking system
Rita Paverman Kashani, Meitar (IL); Shaked Dolev, Tel Aviv (IL); Yuval Snappir, Tel Aviv (IL); Daniel Gabay, Even Yehuda (IL); and Sivan Keret, Givatayim (IL)
Assigned to TRIGO VISION LTD., Tel Aviv (IL)
Filed by TRIGO VISION LTD., Tel Aviv (IL)
Filed on Apr. 18, 2022, as Appl. No. 17/722,970.
Application 17/722,970 is a continuation of application No. 17/769,908, previously published as PCT/IL2020/051198, filed on Nov. 19, 2020.
Claims priority of provisional application 62/938,563, filed on Nov. 21, 2019.
Claims priority of provisional application 62/938,681, filed on Nov. 21, 2019.
Prior Publication US 2022/0335726 A1, Oct. 20, 2022
Int. Cl. G06V 20/52 (2022.01); G06T 7/292 (2017.01); G06T 7/73 (2017.01); G06V 10/22 (2022.01)
CPC G06V 20/52 (2022.01) [G06T 7/292 (2017.01); G06T 7/73 (2017.01); G06V 10/22 (2022.01); G06T 2207/20081 (2013.01)] 16 Claims
OG exemplary drawing
 
1. A method for processing visual data relating to an object, the method comprising:
causing at least one camera to view a scene; and
employing visual analysis for:
ascertaining whether at least one output from said at least one camera does or does not contain information indicating the presence and location of at least one object in said scene, and
enhancing at least one output that does not contain information indicating said presence and location of said at least one object in said scene by adding thereto information indicating said presence and location of said at least one object in said scene, said added information being derived based on outputs from at least two other cameras arrange to view said scene,
said enhancing comprising:
identifying at least one pixel location associated with a bounding box indicating said presence and location of said at least one object in said outputs from said at least two other cameras;
for each said pixel location, finding a camera ray corresponding thereto, for said at least two other cameras;
performing matching between ones of said camera rays to find a point of intersection therebetween, said point of intersection corresponding to said location of said object;
finding a camera ray corresponding to said object location, for said at least one camera having an output not containing information indicating said presence and location of said at least one object;
finding a pixel location corresponding to said camera ray, for said at least one camera having an output not containing information indicating said presence and location of said at least one object; and
inserting a new bounding box in said output from said at least one camera at said pixel location corresponding to said camera ray, said new bounding box indicating said presence and location of said at least one object.