Methods and apparatuses for dropped object detection
AssigneeBoston Dynamics, Inc.
InventorsAndrew Hoelscher, Samuel Shaw, Alexander Ozer, Jennifer Barry, Matthew Turpin
Life of this patent
Estimating the term…
Abstract
Methods and apparatuses for detecting one or more objects (e.g., dropped objects) by a robotic device are described. The method comprises receiving a distance-based point cloud including a plurality of points in three dimensions, filtering the distance-based point cloud to remove points from the plurality of points based on at least one known surface in an environment of the robotic device to produce a filtered distance-based point cloud, clustering points in the filtered distance-based point cloud to produce a set of point clusters, and detecting one or more objects based, at least in part, on the set of point clusters.
Claims23
A method comprising:
- capturing, by at least one first sensor, an image of a stack of objects arranged near a robotic device;
- receiving, by at least one computing device, a distance-based point cloud including a plurality of points in three dimensions, the plurality of points including points corresponding to one or more objects not included in the stack of objects, wherein the distance-based point cloud corresponds to sensor data captured by at least one second sensor arranged to sense objects in multiple directions around a location of the robotic device;
- filtering, by the at least one computing device, the distance-based point cloud to remove points from the plurality of points based on at least one known surface of an environment of the robotic device to produce a filtered distance-based point cloud;
- clustering, by the at least one computing device, points in the filtered distance-based point cloud to produce a set of point clusters;
- detecting, by the at least one computing device, the one or more objects based, at least in part, on the set of point clusters;
- and
- controlling the robotic device to grasp at least one of the detected one or more objects.
A mobile robotic device, comprising:
- at least one first sensor configured to capture an image of a stack of objects arranged near the mobile robotic device;
- at least one second sensor arranged on a base of the robotic device to sense objects in multiple directions around a location of the robotic device, wherein the at least one second sensor is configured to collect a distance-based point cloud including a plurality of points in three dimensions, the plurality of points including points corresponding to one or more objects not included in the stack of objects;
- at least one storage device configured to store information about at least one known surface of an environment of the robotic device;
- an end effector configured to grasp objects in the environment of the robotic device;
- and
- at least one computing device configured to:
- filter the distance-based point cloud to remove points from the plurality of points based on the information about the at least one known surface to produce a filtered distance-based point cloud;
- cluster points in the filtered distance-based point cloud to produce a set of point clusters;
- detect the one or more objects based, at least in part, on the set of point clusters;
- and
- control the end effector to grasp at least one of the detected one or more objects.
A non-transitory computer-readable medium encoded with a plurality of instructions that, when executed by at least one computing device, perform a method comprising:
- receiving, from at least one first sensor, an image of a stack of objects arranged near a robotic device;
- receiving a distance-based point cloud including a plurality of points in three dimensions, the plurality of points including points corresponding to one or more objects not included in the stack of objects, wherein the distance-based point cloud corresponds to sensor data captured by at least one second sensor arranged to sense objects in multiple directions around a location of the robotic device;
- filtering the distance-based point cloud to remove points from the plurality of points based on at least one known surface of an environment of the robotic device to produce a filtered distance-based point cloud;
- clustering points in the filtered distance-based point cloud to produce a set of point clusters;
- detecting the one or more objects based, at least in part, on the set of point clusters;
- and
- controlling the robotic device to grasp at least one of the detected one or more objects.
Dependencies are inferred from the claim text.