Information processing method, information processing device, object detection apparatus, and robot system

1. An information processing method for generating learning data for detecting a position/posture of an object, the method comprising:

receiving design data of the object;

generating object data including shape data of the object and position/posture information of the object from the design data;

generating first bulk data by stacking the object data by simulation;

extracting the object data disposed on an outermost surface as outermost surface object data in the first bulk data;

disposing an imaging section based on an imaging angle and an imaging distance at which a position/posture of the outermost surface object data is obtained, when the object placed alone is imaged by the imaging section;

acquiring first imaging data by imaging the object placed alone by the imaging section from the imaging angle and the imaging distance; and

generating the learning data by replacing the shape data included in the outermost surface object data with the first imaging data, in the first bulk data.

2. The information processing method according to claim 1 , wherein

the design data is data of a three-dimensional CAD of the object.

3. The information processing method according to claim 1 , further comprising:

acquiring fifth imaging data including the shape data of the object by imaging the object placed alone by the imaging section under a plurality of imaging conditions, wherein

the generating the first bulk data is obtaining the first bulk data by associating the object data with the fifth imaging data and stacking the fifth imaging data associated with the object data.

4. An information processing device for generating learning data for detecting a position/posture of an object, the apparatus comprising:

a design data receiving section receiving design data of the object;

an object data generation section generating object data that includes shape data of the object and position/posture information of the object from the design data;

a first bulk data generation section generating first bulk data by stacking the object data by simulation;

an outermost surface extraction section extracting the object data that is disposed on an outermost surface as outermost surface object data, in the first bulk data;

an imaging angle calculation section calculating an imaging angle and an imaging distance corresponding to the outermost surface object data, when the object placed alone is imaged by an imaging section;

a first imaging data acquisition section acquiring first imaging data by imaging the object placed alone by the imaging section at the imaging angle and the imaging distance; and

a learning data generation section generating learning data by replacing the shape data included in the outermost surface object data with the first imaging data, in the first bulk data.

5. A robot system comprising:

a robot including a robot arm;

an imaging section installed in the robot arm;

an object detection apparatus for detecting a position/posture of an object; and

a robot control device controlling driving of the robot based on a detection result of the object detection apparatus, wherein

the object detection apparatus includes:

the information processing device according to claim 4 ;

a second imaging data acquisition section acquiring second imaging data by imaging the object by an imaging section; and

an object position/posture detection section detecting the position/posture of the object based on the second imaging data and the learning data.