Methods and apparatus for determining the pose of an object based on point cloud data
AssigneeX Development LLC
InventorStefan Hinterstoisser
Life of this patent
Estimating the term…
Abstract
Methods, apparatus, and computer readable media that are related to 3D object detection and pose determination and that may optionally increase the robustness and/or efficiency of the 3D object recognition and pose determination. Some implementations are generally directed to techniques for generating an object model of an object based on model point cloud data of the object. Some implementations of the present disclosure are additionally and/or alternatively directed to techniques for application of acquired 3D scene point cloud data to a stored object model of an object to detect the object and/or determine the pose of the object.
Claims21
A method implemented by one or more processors, comprising:
- identifying, by one or more of the processors, scene point cloud data, wherein the scene point cloud data captures at least a portion of an object in an environment;
- identifying, by one or more of the processors, a bounding measure for one or more dimensions of the object;
- selecting, by one or more of the processors, a scene reference point from the scene point cloud data;
- selecting, by one or more of the processors from the scene point cloud data, an additional scene point to pair with the scene reference point as a scene point pair in generating a candidate in-plane rotation, wherein selecting the additional scene point is based on a distance between the scene reference point and the additional scene point satisfying the bounding measure;
- identifying a model point pair based on one or more features of the scene point pair, wherein the identifying is by one or more of the processors from a stored object model of the object;
- generating, by one or more of the processors, the candidate in-plane rotation based on the model point pair and the scene point pair;
- and
- determining, by one or more of the processors, a pose of the object in the environment based on the generated candidate in-plane rotation and a model reference point of the model point pair.
A robot, comprising:
- one or more processors receiving scene point cloud data, the scene point cloud data capturing at least a portion of an object in an environment, wherein the one or more processors are configured to:
- identify a bounding measure for one or more dimensions of the object,
- select a scene reference point from the scene point cloud data;
- select, from the scene point cloud data, an additional scene point to pair with the scene reference point as a scene point pair in generating a candidate in-plane rotation, wherein the one or more processors in selecting the additional scene point are configured to select the additional scene point based on a distance between the scene reference point and the additional scene point satisfying the bounding measure;
- identify, from a stored object model of the object, a model point pair based on one or more features of the scene point pair;
- generate the candidate in-plane rotation for the object based on the model point pair and the scene point pair;
- determine a pose of the object in the environment based on the generated candidate in-plane rotation and a model reference point of the model point pair;
- and
- provide, to one or more actuators of the robot, control commands that are based on the determined pose.
At least one non-transitory computer-readable media comprising instructions that, in response to execution of the instructions by one or more processors, cause the one or more processors to perform the following operations:
- identify scene point cloud data, wherein the scene point cloud data is sensed via one or more three-dimensional vision sensors and captures at least a portion of an object in an environment;
- identify a bounding measure for one or more dimensions of the object,
- select a scene reference point from the scene point cloud data;
- select, from the scene point cloud data, an additional scene point to pair with the scene reference point as a scene point pair in generating a candidate in-plane rotation, wherein selecting the additional scene point is based on a distance between the scene reference point and the additional scene point satisfying the bounding measure;
- identify, from a stored object model of the object, a model point pair based on one or more features of the scene point pair;
- generate the candidate in-plane rotation for the object based on the model point pair and the scene point pair;
- determine a pose of the object in the environment based on the generated candidate in-plane rotation and a model reference point of the model point pair.
Dependencies are inferred from the claim text.