Learned part system

• US 4,704,694 A
• Filed: 12/16/1985
• Issued: 11/03/1987
• Est. Priority Date: 12/16/1985
• Status: Expired due to Term

• Current Assignee: Schneider Automation Incorporated (Schneider Electric SE)
• Original Assignee: Automation Intelligence Incorporated
• Inventors: Czerniejewski, Frank R.
• Primary Examiner(s): Smith, Jerry
• Assistant Examiner(s): MACDONALD, ALLEN R
• Application Number:
  US06/809,208
• Time in Patent Office: 687 Days
• Field of Search: 364/513, 364/513.5, 364/576, 364/478, 901/47, 356/388, 356/389, 356/390-398, 209/576-579, 382/25, 382/43, 358/101
• US Class Current: 382/288
• CPC Class Codes: B25J 9/1697 Vision controlled systems
• CPC Class Codes: G06T 1/0014 Image feed-back for automatic industrial control, e.g. robot with camera robots B25J19/023

Abstract

A system for learning a characterization of an object and thereafter identifying the presence and orientation of the object is capable of controlling a robotic system for grasping the object. In an off-line learning mode, a television camera produces a video signal from the object and a preprocessor develops a segmented line outline of the object in digital form. A computer is programmed to generate a mathematical "rubber band" or convex hull around the object outline formed from segmented lines and to store the x-y coordinates of the lines. The centroid of the hull is calculated and an R(.0.) function as the radius from the centroid to the hull with the maxima thereof is calculated for characterizing the orientation of the object. A truncated Fourier series of the R(.0.) function is generated and the coefficients stored. In an on-line characterization mode, the apparatus generates a characterization of an unknown object and compares it with the learned and stored values of R(.0.) and Fourier coefficients. Upon identification of the object, the orientation is calculated by comparing the shift in .0. from the stored reference values.

Claims

1. A system for characterizing an object in a first orientation thereof, said system having a learning mode and an identifying mode, comprising:

(a) camera means for producing video signals representative of the outline of an object to be characterized in the learning mode and of an object to be identified in the identifying mode;

(b) signal processing means connected to said camera means for detecting the edges of said outline from said video signals and for producing digital signals defining a multiplicity of straight line segments by a first set of x-y coordinates representative of said outline, said processing means including first storage means for storing a table of said first set of x-y coordinates;

(c) computer means connected to said processing means and having first program means for forming a convex hull around said outline by defining a second set of x-y coordinates and calculating characterizing parameters of said object from said second set of x-y coordinates, said parameters including

(i) a centroid of said convex hull,

(ii) an area of said convex hull,

(iii) a radius R of said convex hull as a function of angle φ

with respect to said first orientation of the object thereby producing a single valued mathematical function R(φ

) defining said convex hull, and

(iv) a Fourier series representation of said hull defining function R(φ

) truncated at a preselected number of terms; and

(d) second storage means connected to said computer means for storing said calculated characterizing parameters when said system is in the learning mode;

(e) second program means programmed into said computer means for comparing characterizing parameters produced from another object, when said system is in the identifying mode, to characterizing parameters previously stored in said second storage means and for identifying such previously stored characterizing parameters which are essentially the same as said characterizing parameter of said other object so as to identify such other object; and

(f) third program means programmed into said computer means for calculating the difference between the angle φ

of the hull defining function R(φ

) of said other object and the angle φ

of said identified previously stored characterizing parameters thereby determining the amount of rotation of such other object with respect to the orientation of the object producing such stored parameters during said learning mode.

2. The system as defined in claim 1 in which said third program means further includes means for calculating the difference in the angle of the arguments of the said Fourier series representation of said hull defining function R(φ

) of said other object and the arguments of the stored Fourier series representation of an R(φ

) function of said identified previously stored characterizing parameters thereby determining the amount of rotation of such other object with respect to the orientation of the object producing such stored Fourier series during said learning mode.

3. The system as recited in claim 1 which further comprises:

a video monitor connected to said camera means for displaying said outline produced by said signal processing means, said monitor having a cursor;

cursor control means for moving said cursor to form a window within said outline on said monitor; and

in which said first program means includes means for forming a hull around said window by defining a third set of x-y coordinates and storing said third set of x-y coordinates in said second storage means.

4. In a system having a plurality of objects and a robotic system for handling objects, apparatus for learning objects in an off-line mode, for identifying learned objects in an on-line mode and for controlling said robotic system to grasp such objects comprising:

a television camera for producing an analog video signal of an unknown object within its field of view;

an analog to digital converter connected to said camera for digitizing said analog video signal;

a gradient detector connected to said analog to digital converter for detecting the edges of an outline of said object from said digitized video signal and producing a set of straight line segments representative of said outline;

first storage means connected to said gradient detector for storing coordinates of said straight line segments;

computer means for accessing said first storage means for forming a convex hull around the representation of said outline, said hull being represented as a multiplicity of straight line segments, and for calculating the centroid and the area of said hull;

said computer means comprising means for calculating a function R(φ

) in which R is the radius from said centroid to said hull and φ

is the angle of said radius with respect to a selected axis wherein maxima of said function define the orientation of said object with respect to said axis;

second storage means connected to said computer means for storing said function when said system is in the off-line mode, said computer means comprising means for comparing, when said system is in the on-line mode, said function with functions stored in said second storage means to identify said unknown object and to calculate the orientation of said unknown object with respect to said axis; and

interface means for connecting said computer means to said robotic system for controlling said system to grasp said object.

5. The system as recited in claim 4 in which:

said computer means comprising means for calculating a truncated Fourier series representation of said R(φ

) function; and

said second storage means is also for storing the coefficients and arguments of each term of said calculated Fourier series when said system is in the off-line mode.

6. The system as recited in claim 5 in which said computer means comprising means for comparing the coefficients and arguments of each term of said calculated Fourier series when said system is in the on-line mode with said stored coefficients and arguments to thereby identify said unknown object by said coefficients and the orientation thereof by said arguments.