System and method for object detector training
AssigneeIntrinsic Innovation LLC
InventorsSivaramakrishnan Swaminathan, Eyrun Eyjolfsdottir
Life of this patent
Estimating the term…
Abstract
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for automatically generating object representations. One of the methods includes grasping, by a robot, an object at a first grasp point and generating a first partial object mesh based on one or more first sensor measurements of the object when held by the robot at the first grasp point. A second grasp point is identified for the object that is located in a region captured by the one or more first sensor measurements. A second partial object mesh is generated based on one or more second sensor measurements of the object when held by the robot at the second grasp point.
Claims14
A method comprising:
- identifying, by a sensor associated with a robot, a first grasp point for an object to be automatically modeled;
- grasping, by the robot, the object at the first grasp point;
- obtaining one or more first sensor measurements of the object;
- generating a first partial object mesh based on the one or more first sensor measurements of the object when held by the robot at the first grasp point;
- identifying a second grasp point for the object that is located in a region captured by the one or more first sensor measurements;
- grasping, by the robot, the object at the second grasp point;
- obtaining one or more second sensor measurements for the object;
- generating a second partial object mesh based on the one or more second sensor measurements of the object when held by the robot at the second grasp point;
- generating a final object mesh for the object including aligning and merging the first partial object mesh and the second partial object mesh;
- and
- repeatedly generating additional partial object meshes until determining that a stopping condition is reached, wherein determining that the stopping condition is reached comprises:
- generating an alignment angle from aligning a first partial object mesh and an additional partial object mesh;
- and
- determining that the alignment angle satisfies a threshold.
A system comprising:
- one or more computers and one or more storage devices storing instructions that are operable, when executed by the one or more computers, to cause the one or more computers to perform operations comprising:
- identifying, by a sensor associated with a robot, a first grasp point for an object to be automatically modeled;
- grasping, by the robot, the object at the first grasp point;
- obtaining one or more first sensor measurements of the object;
- generating a first partial object mesh based on the one or more first sensor measurements of the object when held by the robot at the first grasp point;
- identifying a second grasp point for the object that is located in a region captured by the one or more first sensor measurements;
- grasping, by the robot, the object at the second grasp point;
- obtaining one or more second sensor measurements for the object;
- generating a second partial object mesh based on the one or more second sensor measurements of the object when held by the robot at the second grasp point;
- generating a final object mesh for the object including aligning and merging the first partial object mesh and the second partial object mesh;
- and
- repeatedly generating additional partial object meshes until determining that a stopping condition is reached, wherein determining that the stopping condition is reached comprises:
- generating an alignment angle from aligning a first partial object mesh and an additional partial object mesh;
- and
- determining that the alignment angle satisfies a threshold.
One or more non-transitory computer storage media encoded with computer program instructions that when executed by one or more computers cause the one or more computers to perform operations comprising:
- identifying, by a sensor associated with a robot, a first grasp point for an object to be automatically modeled;
- grasping, by the robot, the object at the first grasp point;
- obtaining one or more first sensor measurements of the object;
- generating a first partial object mesh based on the one or more first sensor measurements of the object when held by the robot at the first grasp point;
- identifying a second grasp point for the object that is located in a region captured by the one or more first sensor measurements;
- grasping, by the robot, the object at the second grasp point;
- obtaining one or more second sensor measurements for the object;
- generating a second partial object mesh based on the one or more second sensor measurements of the object when held by the robot at the second grasp point;
- generating a final object mesh for the object including aligning and merging the first partial object mesh and the second partial object mesh;
- and
- repeatedly generating additional partial object meshes until determining that a stopping condition is reached, wherein determining that the stopping condition is reached comprises:
- generating an alignment angle from aligning a first partial object mesh and an additional partial object mesh;
- and
- determining that the alignment angle satisfies a threshold.
Dependencies are inferred from the claim text.