Method and apparatus for single image 3D vision guided robotics
DCFirst Claim
Patent Images
1. A useful in three-dimensional pose estimation for use with a single camera mounted to a movable portion of a robot, the method comprising:
- capturing a two-dimensional image of a volume containing a target object;
locating a number of features in the captured image of the target object; and
determining by a processor an object space-to-camera space transformation for the target object based at least in part on a position of at least some of the located features using only the single captured image and an algorithm that employs a known or determinable physical relationship between at least some of the located features.
5 Assignments
Litigations
1 Petition
Accused Products
Abstract
A method of three-dimensional object location and guidance to allow robotic manipulation of an object with variable position and orientation using a sensor array which is a collection of one or more sensors capable of forming a single image.
219 Citations
28 Claims
-
1. A useful in three-dimensional pose estimation for use with a single camera mounted to a movable portion of a robot, the method comprising:
-
capturing a two-dimensional image of a volume containing a target object; locating a number of features in the captured image of the target object; and determining by a processor an object space-to-camera space transformation for the target object based at least in part on a position of at least some of the located features using only the single captured image and an algorithm that employs a known or determinable physical relationship between at least some of the located features. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19)
-
-
20. An apparatus useful in robotics, the apparatus comprising:
-
a single camera operable to capture a number of images of a calibration object; means for calibrating the camera, by; determining a set of intrinsic parameters of the camera from at least one of the number of images of the calibration object captured by the camera; and determining a set of extrinsic parameters of the camera from at least one of the number of images of the calibration object captured by the camera, the set of extrinsic parameters comprising a camera space-to-training space transformation defining a transformation between a camera space reference frame and a training space reference frame; and means for estimating a pose of a target object, by; capturing a two-dimensional image of a volume containing a target object; and locating at least six features in the captured image of the target object; and determining an object space-to-camera space transformation based at least in part on a position of at least some of the located features in solely the captured image using an algorithm that employs a known or determinable physical relationship between at least some of the located features. - View Dependent Claims (21, 22, 23, 24)
-
-
25. An apparatus useful in robotics, the apparatus comprising:
-
a single camera operable to capture a number of images of a calibration object; means for calibrating the camera, by; determining a set of intrinsic parameters of the camera from at least one of the number of images of the calibration object captured by the camera; and determining a set of extrinsic parameters of the camera from at least one of the number of images of the calibration object captured by the camera, the set of extrinsic parameters comprising a camera space-to-training space transformation defining a transformation between a camera space reference frame and a training space reference frame; and means for estimating a pose of a target object, by; capturing a two-dimensional image of a volume containing a target object; locating at least five features in the captured image of the target object; and determining an object space-to-camera space transformation based at least in part on a position of at least some of the located features using the captured image without any additional captured images and an algorithm that employs a known or determinable physical relationship between at least some of the located features. - View Dependent Claims (26, 27, 28)
-
Specification