US 12,272,109 B2
Method and system for automated debris detection
Giacomo Vianello, Mountain View, CA (US); Robert Davis, Mountain View, CA (US); John K. Clark, Mountain View, CA (US); and Jonathan M. Fisher, Mountain View, CA (US)
Assigned to Cape Analytics, Inc., Palo Alto, CA (US)
Filed by Cape Analytics, Inc., Mountain View, CA (US)
Filed on May 20, 2022, as Appl. No. 17/749,385.
Application 17/749,385 is a continuation of application No. 17/502,825, filed on Oct. 15, 2021, granted, now 11,367,265.
Claims priority of provisional application 63/092,283, filed on Oct. 15, 2020.
Prior Publication US 2022/0277538 A1, Sep. 1, 2022
This patent is subject to a terminal disclaimer.
Int. Cl. G06K 9/00 (2022.01); G06F 18/24 (2023.01); G06F 30/13 (2020.01); G06T 7/50 (2017.01); G06V 10/22 (2022.01); G06V 10/25 (2022.01); G06V 10/26 (2022.01); G06V 20/10 (2022.01); G06N 20/00 (2019.01)
CPC G06V 10/22 (2022.01) [G06F 18/24 (2023.01); G06F 30/13 (2020.01); G06T 7/50 (2017.01); G06V 10/25 (2022.01); G06V 10/273 (2022.01); G06V 20/176 (2022.01); G06V 20/188 (2022.01); G06N 20/00 (2019.01)] 19 Claims
OG exemplary drawing
 
1. A method for automatic debris detection comprising:
determining a set of region measurements depicting a set of properties, wherein the region measurement comprises depth information;
determining a foreground representation using the depth information to differentiate terrain from foreground objects;
determining a non-debris representation based on the set of region measurements, wherein determining the non-debris representation comprises identifying instances of a property feature within the set of region measurements;
identifying debris for a property of the set of properties based on the foreground representation and the non-debris representation; and
generating a debris score for the property based on the identified debris.