Object pickup device and method for picking up object
AssigneeFanuc Corporation
InventorKazunori Ban
Life of this patent
Estimating the term…
Abstract
A pickup device for picking up a target object from a plurality of objects randomly piled up in a container, and for placing the target object in a predetermined posture to a target location is provided. The device includes an approximate position obtaining part for obtaining information on an approximate position of the target object, based on information on a height distribution of the objects in the container, which is obtained by a first visual sensor. The device also includes a placement operation controlling part for controlling a robot so as to bring the target object into a predetermined position and posture relative to the target location, based on information on a position and posture of the target object relative to a robot, which is obtained by a second visual sensor.
Claims13
A pickup device configured to pick up by using a robot a target object from a plurality of objects randomly piled up in a container and configured to place the target object in a predetermined posture at a target location, wherein the robot comprises a holding unit configured to hold the target object to take the target object out of the container,
- the pickup device comprising:
- a first visual sensor configured to compute a height distribution of a visible subset of the plurality of objects in the container visible to the first visual sensor, the height distribution including the calculated heights for each of the objects in the visible subset;
- an approximate position obtaining part configured to obtain information on an approximate position of the target object which is at a higher position than surrounding objects in the visible subset, based on the information on a height distribution obtained by the first visual sensor and exclusive of a posture of the target object;
- a pickup operation controlling part configured to control the robot so as to allow the holding unit to hold the target object, based on the information on an approximate position of the target object obtained by the approximate position obtaining part and exclusive of the posture of the target object;
- a second visual sensor configured to obtain information on a position, and the posture of the target object relative to the holding unit by measuring the target object when the target object is held by the holding unit;
- and
- a placement operation controlling part configured to control the robot so as to bring the target object held by the holding unit into a predetermined position and the predetermined posture relative to the target location, based on the position and the posture of the target object relative to the holding unit obtained by the second visual sensor.
A method configured to pick up by using a robot a target object from a plurality of objects randomly piled up in a container and configured to place the target object in a predetermined posture at a target location, wherein the robot comprises a holding unit configured to hold the target object to take the target object out of the container,
- the method comprising:
- computing a height distribution of a visual subset of the plurality of objects in the container visible to a first visual sensor, the height distribution including the calculated heights for each of the objects in the visible subset;
- obtaining information on an approximate position of the target object which is at a higher position than surrounding objects in the visible subset, based on the information on a height distribution obtained by the first visual sensor and exclusive of a posture of the target object;
- controlling the robot so as to allow the holding unit to hold the target object, based on the information on an approximate position of the target object and exclusive of the posture of the target object;
- obtaining information on a position and the posture of the target object relative to the holding unit by a second visual sensor, by measuring the target object in a state where the target object is held by the holding unit;
- and
- controlling the robot so as to bring the target object held by the holding unit into a predetermined position and the predetermined posture relative to the target location, based on the information on a position and the posture of the target object relative to the holding unit obtained by the second visual sensor.
Dependencies are inferred from the claim text.