IP Library Granted Patent US 12,654,312
Granted Patent B2
US 12,654,312 · App. 17/959,009 · Granted Jun 16, 2026

Systems and methods for assisting in object recognition in object processing systems

Inventors: Luis Galup (Sudbury, MA); Thomas Koletschka (Natick, MA)
Assignee: Berkshire Grey Operating Company, Inc.
B25J9/1612B25J9/0093B25J15/0616B65G47/91G06T1/0014G06T7/74G06V20/50
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 12,654,312
App. No.
17/959,009
Granted
Jun 16, 2026
Kind
B2
Abstract

An object recognition system includes: an image capture system for capturing at least one image of an object, and for providing image data representative of the captured image; a patch identification system in communication with the image capture system for receiving the image data, and for identifying at least one image patch associated with the captured image, each image patch being associated with a potential grasp location on the object, each potential grasp location being described as an area that may be associated with a contact portion of an end effector of a programmable motion device; a feature identification system for capturing at least one feature of each image patch, for accessing feature image data in the database and for providing feature identification data responsive to the image feature comparison data; and an object identification system for providing object identify data responsive to the image feature comparison data.

Claims (45)

1 . An object recognition system in communication with a database, said object recognition system comprising:

an image capture system for capturing at least one image of an object, and for providing image data representative of the captured image;

a patch identification system in communication with the image capture system for receiving the image data, and for identifying more than one image patch associated with the captured image, each image patch being associated with a potential grasp location on the object, each potential grasp location being described as an area corresponding to a contact portion of an end effector of a programmable motion device;

a feature identification system for capturing at least one feature of each image patch, the at least one feature of each image patch being a unique pattern exhibited by neighboring pixels within the image patch, for accessing feature image data in the database against which each captured at least one feature is compared to provide image feature comparison data, and for providing feature identification data responsive to the image feature comparison data; and

an object identification system for providing object identity data responsive to the image feature comparison data, said object identity data including data representative of an identity of the object as well as at least one grasp location on the object.

2 . The object recognition system as claimed in claim 1 , wherein the contact portion of the end effector of the programmable motion device is a contact portion of a vacuum cup end effector.

3 . The object recognition system as claimed in claim 1 , wherein the contact portion of the end effector of the programmable motion device is an object-level patch that includes a contact portion of a vacuum cup end effector.

4 . The object recognition system as claimed in claim 1 , wherein the object identification system provides patch center data representative of a center of each patch, and the patch center data is provided responsive to the at least one captured feature.

5 . The object recognition system as claimed in claim 1 , wherein the object recognition system further includes an object image identification system in communication with the image capture system for receiving the image data, and for identifying an object associated with at least one object image associated with the captured image.

6 . The object recognition system as claimed in claim 1 , wherein the system prioritizes patch matches over model-free grasps.

7 . The object recognition system as claimed in claim 1 , wherein the system includes an input station at which objects are provided in input containers.

8 . The object recognition system as claimed in claim 1 , wherein the system includes an input station at which objects are provided on an input conveyor.

9 . The object recognition system as claimed in claim 1 , wherein the object recognition system further includes an additional scanner through which objects may be passed by the programmable motion device.

10 . The object recognition system as claimed in claim 1 , wherein the object recognition system further includes an output station that includes at least one shuttle wing sortation system.

11 . An object processing system in communication with a database, said object processing system comprising:

an input conveyor providing an input station at which a plurality of objects may be received;

a plurality of containers at an output station at which the plurality of objects may be provided among a plurality of destination locations; and

a programmable motion device operating at a processing station intermediate the input station and the output station, said processing station including:

an image capture system for capturing at least one image of an object of the plurality of objects, and for providing image data representative of the captured image;

a patch identification system in communication with the image capture system for receiving the image data, and for identifying more than one image patch associated with the captured image, each image patch being associated with a potential grasp location on the object;

a feature identification system for capturing at least one feature of each image patch, the at least one feature being a unique pattern in the image patch exhibited by neighboring pixels, for accessing feature image data in the database against which each captured at least one feature is compared to provide image feature comparison data; and

an object identification system for providing object identity data responsive to the image feature comparison data, said object identity data including data representative of an identity of the object as well as at least one grasp location on the object.

12 . The object processing system as claimed in claim 11 , wherein each potential grasp location is described as an area that may be associated with a contact portion of the end effector of the programmable motion device.

13 . The object processing system as claimed in claim 12 , wherein the contact portion of the end effector of the programmable motion device is a contact portion of a vacuum cup end effector.

14 . The object processing system as claimed in claim 12 , wherein the contact portion of the end effector of the programmable motion device is an object-level patch that includes a contact portion of a vacuum cup end effector.

15 . The object processing system as claimed in claim 11 , wherein the object identification system provides patch center data representative of a center of each patch, and the patch center data is provided responsive to the at least one captured feature.

16 . The object processing system as claimed in claim 11 , wherein the object recognition system further includes an object image identification system in communication with the image capture system for receiving the image data, and for identifying an object associated with at least one object image associated with the captured image.

17 . The object processing system as claimed in claim 11 , wherein the system prioritizes patch matches over model-free grasps.

18 . The object processing system as claimed in claim 11 , wherein the system includes an input station at which objects are provided on an input conveyor that is fed by a cleated conveyor.

19 . The object processing system as claimed in claim 11 , wherein the object recognition system further includes a drop scanner through which objects may be dropped by the programmable motion device.

20 . The object processing system as claimed in claim 11 , wherein the output station includes at least one shuttle wing sortation system.

21 . The object processing system as claimed in claim 11 , wherein the input station includes a recirculating conveyor for recirculating objects to the processing station.

22 . The object processing system as claimed in claim 21 , wherein the input station further includes a feeding conveyor for providing objects from an input hopper to the input station.

23 . A method of providing object recognition comprising:

capturing at least one image of an object, and for providing image data representative of the captured image;

receiving the image data, and for identifying more than one image patch associated with the captured image, each image patch being associated with a potential grasp location on the object, each potential grasp location being described as an area corresponding to a contact portion of an end effector of a programmable motion device;

capturing at least one feature of each image patch, the at least one feature of each image patch being a unique pattern exhibited by neighboring pixels within the image patch, for accessing feature image data in the database against which each captured at least one feature is compared to provide image feature comparison data, and for providing feature identification data responsive to the image feature comparison data; and

providing object identity data responsive to the image feature comparison data, said object identity data including data representative of an identity of the object as well as at least one grasp location on the object.

24 . The method as claimed in claim 23 , wherein the contact portion of the end effector of the programmable motion device is a contact portion of a vacuum cup end effector.

25 . The method as claimed in claim 23 , wherein the contact portion of the end effector of the programmable motion device is an object-level patch that includes a contact portion of a vacuum cup end effector.

26 . The method as claimed in claim 23 , wherein the object identification system provides patch center data representative of a center of each patch, and the patch center data is provided responsive to the at least one captured feature.

27 . The method as claimed in claim 23 , wherein the method further includes identifying an object associated with at least one object image associated with the captured image.

28 . The method as claimed in claim 23 , wherein the method further includes prioritizing patch matches over model-free grasps.

29 . The method as claimed in claim 23 , wherein the method further includes dropping objects through a drop scanner onto an output conveyance system.

30 . The method as claimed in claim 23 , wherein the method further includes dropping objects from a reciprocating carriage into any of a plurality of destination locations based on the object identity data.

Assignments (1)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jan 25, 2023
From: GALUP, LUIS; KOLETSCHKA, THOMAS
To: BERKSHIRE GREY OPERATING COMPANY, INC.
Reel/Frame 062476/0122 →
Continuity (2)
Provisional Application 63256397 · Oct 15, 2021
Related Publication 20230124854A1 · Apr 20, 2023
References Cited (108)
US 4557659A · Scaglia · 1985 [cited by applicant]
US 4786847A · Daggett et al. · 1988 [cited by applicant]
US 4923067A · Fuller, Jr · 1990 [cited by examiner]
US 5495097A · Katz · 1996 [cited by examiner]
US 5865487A · Gore et al. · 1999 [cited by applicant]
US 5937143A · Watanabe et al. · 1999 [cited by applicant]
US 6060677A · Ulrichsen et al. · 2000 [cited by applicant]
US 6167607B1 · Pryor · 2001 [cited by applicant]
US 6721444B1 · Gu et al. · 2004 [cited by applicant]
US 7313464B1 · Perreault et al. · 2007 [cited by applicant]
US 7474939B2 · Oda et al. · 2009 [cited by applicant]
US 8132835B2 · Ban et al. · 2012 [cited by applicant]
US 9102055B1 · Konolige et al. · 2015 [cited by applicant]
US 9266237B2 · Nomura · 2016 [cited by applicant]
US 9283680B2 · Yasuda et al. · 2016 [cited by applicant]
US 9492923B2 · Wellman et al. · 2016 [cited by applicant]
US 9604363B2 · Ban · 2017 [cited by applicant]
US 9687983B1 · Prats · 2017 [cited by applicant]
US 9926138B1 · Brazeau et al. · 2018 [cited by applicant]
US 9937532B2 · Wagner · 2018 [cited by examiner]
US 10007827B2 · Wagner et al. · 2018 [cited by applicant]
US 10369701B1 · Diankov et al. · 2019 [cited by applicant]
US 10456915B1 · Diankov · 2019 [cited by applicant]
US 10583986B2 · Wagner · 2020 [cited by examiner]
US 10625305B2 · Wagner · 2020 [cited by examiner]
US 10625432B2 · Wagner · 2020 [cited by examiner]
US 10758943B1 · Carpenter · 2020 [cited by examiner]
US 10913615B2 · Wagner · 2021 [cited by examiner]
US 11420329B2 · Wagner et al. · 2022 [cited by applicant]
US 11813744B2 · Wagner · 2023 [cited by examiner]
US 12002240B2 · Taylor · 2024 [cited by examiner]
US 20010056313A1 · Osborne, Jr. · 2001 [cited by applicant]
US 20020147568A1 · Wenzel et al. · 2002 [cited by applicant]
US 20080181485A1 · Beis et al. · 2008 [cited by applicant]
US 20130054029A1 · Huang et al. · 2013 [cited by applicant]
US 20130266205A1 · Valpola · 2013 [cited by examiner]
US 20130343640A1 · Buehler et al. · 2013 [cited by applicant]
US 20130346348A1 · Buehler · 2013 [cited by examiner]
US 20140067127A1 · Gotou · 2014 [cited by applicant]
US 20150081090A1 · Dong · 2015 [cited by applicant]
US 20150224650A1 · Xu et al. · 2015 [cited by applicant]
US 20160167227A1 · Wellman et al. · 2016 [cited by applicant]
US 20160221187A1 · Bradski et al. · 2016 [cited by applicant]
US 20160245641A1 · Kowdle · 2016 [cited by examiner]
US 20160379076A1 · Nobuoka et al. · 2016 [cited by applicant]
US 20170021499A1 · Wellman et al. · 2017 [cited by applicant]
US 20170050315A1 · Henry et al. · 2017 [cited by applicant]
US 20170121113A1 · Wagner · 2017 [cited by examiner]
US 20170136632A1 · Wagner et al. · 2017 [cited by applicant]
US 20170173638A1 · Wagner · 2017 [cited by examiner]
US 20170225330A1 · Wagner et al. · 2017 [cited by applicant]
US 20180104829A1 · Altman · 2018 [cited by examiner]
US 20180273295A1 · Wagner · 2018 [cited by examiner]
US 20180330134A1 · Wagner · 2018 [cited by examiner]
US 20190002202A1 · Collin · 2019 [cited by examiner]
US 20190050670A1 · Pogorelik · 2019 [cited by examiner]
US 20190061174A1 · Robinson · 2019 [cited by examiner]
US 20190099891A1 · Tomioka · 2019 [cited by examiner]
US 20190337152A1 · Homberg · 2019 [cited by examiner]
US 20190344447A1 · Wicks · 2019 [cited by examiner]
US 20200130935A1 · Wagner et al. · 2020 [cited by applicant]
US 20200143127A1 · Wagner · 2020 [cited by examiner]
US 20200151408A1 · Wagner · 2020 [cited by examiner]
US 20200151409A1 · Wagner et al. · 2020 [cited by applicant]
US 20200151410A1 · Wagner · 2020 [cited by examiner]
US 20200164505A1 · Boer et al. · 2020 [cited by applicant]
US 20200164506A1 · Hallock et al. · 2020 [cited by applicant]
US 20200164531A1 · Wagner · 2020 [cited by examiner]
US 20200189105A1 · Wen et al. · 2020 [cited by applicant]
US 20200242544A1 · Galluzzo et al. · 2020 [cited by applicant]
US 20200319627A1 · Edwards et al. · 2020 [cited by applicant]
US 20200331144A1 · Huang et al. · 2020 [cited by applicant]
US 20210053216A1 · Diankov et al. · 2021 [cited by applicant]
US 20210081791A1 · Goodrich et al. · 2021 [cited by applicant]
US 20210122586A1 · Sun et al. · 2021 [cited by applicant]
US 20210268659A1 · Olmstead · 2021 [cited by examiner]
US 20210276197A1 · Ye · 2021 [cited by examiner]
US 20220261738A1 · Kumar et al. · 2022 [cited by applicant]
US 20220281114A1 · Moreira Rodrigues et al. · 2022 [cited by applicant]
US 20220289502A1 · Sun et al. · 2022 [cited by applicant]
US 20220314447A1 · Wagner et al. · 2022 [cited by applicant]
US 20230071384A1 · Agarwal · 2023 [cited by examiner]
US 20230364787A1 · Cygan · 2023 [cited by examiner]
US 20240149460A1 · Matl · 2024 [cited by examiner]
CA 2928645A1 · 2015 [cited by applicant]
CA 2928645C · 2021 [cited by examiner]
CN 103464383A · 2013 [cited by applicant]
CN 102448679B · 2014 [cited by applicant]
CN 102922521B · 2015 [cited by applicant]
CN 103963058B · 2016 [cited by applicant]
CN 118103880A · 2024 [cited by applicant]
DE 3810989A1 · 1989 [cited by applicant]
DE 10319253A1 · 2004 [cited by applicant]
DE 102007028680A1 · 2008 [cited by applicant]
DE 102007038834A1 · 2009 [cited by applicant]
DE 102010002317A1 · 2011 [cited by applicant]
DE 102012102333A1 · 2013 [cited by applicant]
DE 102018126310B3 · 2019 [cited by examiner]
EP 0317020B1 · 1995 [cited by applicant]
EP 1832947A2 · 2007 [cited by examiner]
JP 6948033B1 · 2021 [cited by examiner]
WO WO2020142296A1 · 2020 [cited by examiner]
WO 2023064134A1 · 2023 [cited by applicant]
Communication pursuant to Rules 161(1) and 162 EPC issued by the European Patent Office in related European Patent Application No. 22793988.1 on May 23, 2024, 3 pages. [cited by applicant]
Notification of Transmittal of the International Search Report and the Written Opinion of the International Searching Authority and the International Search Report and Written Opinion issued by the European Patent Offic… [cited by applicant]
Zeng et al., Robotic Pick-and-Place of Novel Objects in Clutter with Multi-Affordance Grasping and Cross- Domain Image Matching, arxiv.org, Cornell University Library, 201 Olin Library Cornell University Ithaca, NY 1485… [cited by applicant]
Tareen et al., A Comparative Analysis of SIFT, SURF, KAZE, AKAZE, ORB and BRISK, IEEE 2018 International Conference on Computing, Mathematics and Engineering Technologies, iCoMET 2018, 10 pages. [cited by applicant]
International Preliminary Report on Patentability issued by the International Bureau of WIPO in related international application No. PCT/US2022/045556 on Apr. 16, 2024, 7 pages. [cited by applicant]