Back to results
US 6,466,843 B1Granted
Method and apparatus for lifting objects
AssigneeGeneral Electric Company
InventorsPierino Gianni Bonanni, Kevin George Harding, Appasaheb Ningappa Madiwale, Juan Manuel De Bedout
Life of this patent
Estimating the term…
Abstract
In one embodiment of the present invention, a system for lifting objects comprises a Cartesian manipulator, a machine vision system for acquiring video images of the objects, and a control processor for processing the video images and providing position command signals to the Cartesian manipulator.
Claims112
A system for lifting objects comprising:
- a robotic manipulator;
- a distance sensor adapted to acquire object height measurements of said objects;
- a light plane projector adapted to illuminate said objects;
- a video camera adapted to acquire video images of said objects;
- and
- a control processor adapted to process said video images and said object height measurements and provide control signals to said robotic manipulator by implementing a polygonal object location method to provide position command signals to said Cartesian manipulator,
- said polygonal object location method comprising:
- acquiring a video image of said polygonal object using a machine vision system, at least one component of said machine vision system being disposed on a component carrying link of a robotic manipulator;
- locating, in said video image, a new edge point corresponding to a respective edge of said polygonal object;
- incorporating said new edge point into a historical point set;
- computing a plurality of edge lines from said historical point set;
- computing a polygon center from said edge lines;
- and
- commanding said robotic manipulator to move an end effector toward said polygon center by moving said component carrying link.
11 more elements
A system for lifting objects comprising:
- a gantry crane;
- a distance sensor disposed on a trolley of said gantry crane;
- a light plane projector disposed on said trolley of said gantry crane;
- a video camera disposed on a bridge of said gantry crane;
- and
- a control processor adapted to process video images from said video camera and object height measurements from said distance sensor and provide control signals to said gantry crane by implementing a polygonal object location method to provide position command signals to said Cartesian manipulator,
- said polygonal object location method comprising:
- acquiring a video image of said polygonal object using a machine vision system, at least one component of said machine vision system being disposed on a component carrying link of a robotic manipulator;
- locating, in said video image, a new edge point corresponding to a respective edge of said polygonal object;
- incorporating said new edge point into a historical point set;
- computing a plurality of edge lines from said historical point set;
- computing a polygon center from said edge lines;
- and
- commanding said robotic manipulator to move an end effector toward said polygon center by moving said component carrying link.
11 more elements
A method of lifting objects comprising:
- projecting a light plane to illuminate said objects;
- acquiring video images of said objects;
- providing object height measurements of said objects;
- and
- processing said video images and said object height measurements using a polygonal object location algorithm to provide position command signals to a robotic manipulator,
- said polygonal object location method comprising:
- acquiring a video image of said polygonal object using a machine vision system, at least one component of said machine vision system being disposed on a component carrying link of a robotic manipulator;
- locating, in said video image, a new edge point corresponding to a respective edge of said polygonal object;
- incorporating said new edge point into a historical point set;
- computing a plurality of edge lines from said historical point set;
- computing a polygon center from said edge lines;
- and
- commanding said robotic manipulator to move an end effector toward said polygon center by moving said component carrying link.
10 more elements
A method of lifting objects comprising:
- projecting a light plane to illuminate said objects using a light plane projector;
- acquiring video images of said objects using a video camera;
- providing object height measurements of said objects;
- and
- processing said video images and said object height measurements using a polygonal object location algorithm to provide position command signals to a gantry crane,
- said light plane projector being disposed on a trolley of said gantry crane,
- said video camera being disposed on a bridge of said gantry crane,
- said polygonal object location method comprising:
- acquiring a video image of said polygonal object using a machine vision system, at least one component of said machine vision system being disposed on a component carrying link of a robotic manipulator;
- locating, in said video image, a new edge point corresponding to a respective edge of said polygonal object;
- incorporating said new edge point into a historical point set;
- computing a plurality of edge lines from said historical point set;
- computing a polygon center from said edge lines;
- and
- commanding said robotic manipulator to move an end effector toward said polygon center by moving said component carrying link.
12 more elements
A method of locating a polygonal object comprising:
- acquiring a video image of said polygonal object using a machine vision system, at least one component of said machine vision system being disposed on a component carrying link of a robotic manipulator;
- locating, in said video image, a new edge point corresponding to a respective edge of said polygonal object;
- incorporating said new edge point into a historical point set;
- computing a plurality of edge lines from said historical. point set;
- computing a polygon center from said edge lines;
- and
- commanding said robotic manipulator to move an end effector toward said polygon center by moving said component carrying link.
4 more elements
A projection-based method of locating a polygonal object comprising:
- acquiring a video image of said polygonal object using a machine vision system, at least one component of said machine vision system being disposed on a component carrying link of a robotic manipulator;
- projecting a light plane to intersect a horizontal face of said polygonal object whereby said video image comprises at least one structured light line;
- computing, from said video image, a projection array having a plurality of projection array rows and a plurality of projection array columns, said projection array columns respectively corresponding to a plurality of projection angles and
- each comprising a plurality of column elements;
- computing a plurality of angular performance values from said projection array columns;
- finding a maximum performance value among said angular performance values, said maximum performance value corresponding to a critical angle;
- computing, from said video image, a plurality of critical angle projection values at said critical angle;
- finding, among said critical angle projection values, at least one relative maximum projection value respectively corresponding to at least one relative maximum offset;
- selecting a critical offset among said at least one relative maximum offset;
- masking said video image in a neighborhood of a line defined by said critical angle and said critical offset to yield a masked image;
- computing a plurality of orthogonal projection values -in a direction orthogonal to said critical angle;
- detecting a plurality of line segment offsets whereat said orthogonal projection values cross respective ones of a plurality of end point threshold values in a direction defined by a reference ray;
- selecting a line segment endpoint among said line segment offsets;
- locating a new edge point from said critical angle, said critical offset, and said line segment endpoint;
- incorporating said new edge point into a historical point set;
- computing a plurality of edge lines from said historical point set;
- computing a polygon center from said edge lines;
- and
- commanding said robotic manipulator to move an end effector toward said polygon center by moving said component carrying link.
16 more elements
An apparatus for locating a polygonal object comprising:
- a machine vision system for acquiring a video image of said polygonal object, at least one component of said machine vision system being disposed on a component carrying link of a robotic manipulator;
- an edge point locator for locating, in said video image, a new edge point corresponding to a respective edge of said polygonal object;
- an edge point accumulator for incorporating said new edge point into a historical point set;
- an edge line computer for computing a plurality of edge lines from said historical point set;
- a polygon center computer for computing a polygon center from said edge lines;
- and
- a manipulator command generator for commanding said robotic manipulator to move an end effector toward said polygon center by moving 'said component carrying link.
4 more elements
An apparatus for locating a polygonal object comprising:
- a machine vision system for acquiring a video image of said polygonal object, at least one component of said machine vision system being disposed on a component carrying link of a robotic manipulator;
- a light plane projector for projecting a light plane to intersect a horizontal face of said polygonal object whereby said video image comprises at least one structured light line;
- a projection array computer for computing, from said video image, a projection array having a plurality of projection array rows and a plurality of projection array columns, said projection array columns respectively corresponding to a plurality of projection angles and each comprising a plurality of column elements;
- an angular performance evaluator for computing a plurality of angular performance values from said projection array columns;
- a maximum finder for finding a maximum performance value among said angular performance values, said maximum performance value corresponding to a critical angle;
- a projection value computer for computing, from said video image, a plurality of critical angle projection values at said critical angle;
- a maximum finder for finding, among said critical angle projection values, at least one relative maximum projection value respectively corresponding to at least one relative maximum offset;
- a maximum selector for selecting a critical offset among said at least one relative maximum offset;
- an image masker for masking said video image in a neighborhood of a line defined by said critical angle and said critical offset to yield a masked image;
- an orthogonal projection computer for computing a plurality of orthogonal projection values in a direction orthogonal to said critical angle;
- a threshold crossing detector for detecting a plurality of line segment offsets whereat said orthogonal projection values cross respective ones of a plurality of end point threshold values in a direction defined by a reference ray;
- an end point selector for selecting a line segment endpoint among said line segment offsets;
- an edge point computer for locating a new edge point from said critical angle, said critical offset, and said line segment endpoint;
- an edge point accumulator for incorporating said new edge point into a historical point set;
- an edge line computer for computing a plurality of edge lines from said historical point set;
- a polygon center computer for computing a polygon center from said edge lines;
- and
- a manipulator command generator for commanding said robotic manipulator to move an end effector toward said polygon center by moving said component carrying link.
15 more elements
A method of detecting a tilted object comprising:
- providing a reference ray having a reference angle;
- estimating a critical angle from a video image;
- differencing said critical angle from said reference angle of said reference ray to yield an error angle;
- and
- comparing an absolute value of said error angle to an angle error tolerance.
2 more elements
A method of detecting a tilted object comprising:
- providing a reference ray having a reference angle;
- computing, from said video image, a projection array having a plurality of projection array rows and a plurality of projection array columns, said projection array columns respectively corresponding to a plurality of projection angles and each comprising a plurality of column elements;
- computing a plurality of angular performance values from said projection array columns;
- and
- finding a maximum performance value among said angular performance values, said maximum performance value corresponding to a critical angle and thereby estimating said critical angle;
- differencing said critical angle from said reference angle of said reference ray to yield an error angle;
- and
- comparing an absolute value of said error angle to an angle error tolerance.
5 more elements
An apparatus for detecting a tilted object comprising:
- a critical angle estimator for estimating a critical angle from a video image;
- an angle differencer for differencing said critical angle from a reference angle of a reference ray to yield an error angle;
- and
- an angle tolerance comparator for comparing an absolute value of said error angle to an angle error tolerance.
1 more element
An apparatus for detecting a tilted object comprising:
- a projection array computer for computing, from said video image, a projection array having a plurality of projection array rows and a plurality of projection array columns, said projection array columns respectively corresponding to a plurality of projection angles and each comprising a plurality of column elements;
- an angular performance evaluator for computing a plurality of angular performance values from said projection array columns;
- and
- a maximum finder for finding a maximum performance value among said angular performance values, said maximum performance value corresponding to a critical angle thereby estimating said critical angle;
- an angle differencer for differencing said critical angle from a reference angle of a reference ray to yield an error angle;
- and
- an angle tolerance comparator for comparing an absolute value of said error angle to an angle error tolerance.
4 more elements
A system for lifting objects comprising:
- a Cartesian manipulator;
- a machine vision system adapted to acquire video images of said objects;
- and
- a control processor adapted to process said video images and provide position command signals to said Cartesian manipulator,
- said machine vision system comprising:
- a structured light projector for illuminating said objects and
- a video camera adapted to acquire said video images of said illuminated objects,
- said structured light projector being disposed on a first link of said Cartesian manipulator,
- said video camera being disposed on a second link of said Cartesian manipulator,
- said first link being different from said second link.
7 more elements
A system for lifting objects comprising:
- a Cartesian manipulator;
- a machine vision system adapted to acquire video images of said objects;
- and
- a control processor adapted to process said video images and provide position command signals to said Cartesian manipulator,
- said machine vision system comprising:
- a structured light projector for illuminating said objects and
- a video camera adapted to acquire said video images of said illuminated objects,
- said video camera comprising an optical bandpass filter having a wavelength pass band,
- said structured light projector comprising a monochromatic light source having a wavelength within said wavelength pass band,
- said wavelength pass band beginning at about 520 nanometers and ending at about 540 nanometers.
7 more elements
A method of lifting objects comprising:
- acquiring video images of said objects using a video camera;
- processing said video images so as to provide position command signals to a Cartesian manipulator;
- projecting structured light onto said objects using a structured light projector,
- said structured light projector being disposed on a first link of said Cartesian manipulator,
- said video camera being disposed on a second link of said Cartesian manipulator,
- said first link being different from said second link.
3 more elements
A method of lifting objects comprising:
- acquiring video images of said objects using a video camera;
- processing said. video images so as to provide position command signals to a Cartesian manipulator;
- projecting structured light onto said objects using a structured light projector, said step of acquiring said video images comprising optical bandpass filtering using an optical filter having a wavelength pass band,
- said step of projecting said structured light comprising producing a monochromatic light having a wavelength within said wavelength pass band,
- said wavelength pass band beginning at about 520 nanometers and ending at about 540 nanometers.
2 more elements
Dependencies are inferred from the claim text.