JP2018075660A - Work-piece gripping position instruction system and picking device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To accurately pick multiple kinds of work-pieces by using a two-dimensional displacement sensor and providing a controller of a robot arm with picking data including depth information.SOLUTION: A work-piece picking device 1 includes: a two-dimensional displacement sensor 3 for scanning a work-piece aggregate stacked on an XY plane 4 at random; a robot arm 5 with gripping claws 6 for gripping one work-piece 7; a controller 8 for controlling the robot arm 5; and a gripping position instruction system 2 for instructing the controller 8 on a gripping position of a work-piece 7 on the basis of waveform data W from the sensor 3. A computer 10 of the system 2 inputs the waveform data W showing a work-piece shape in the XY plane from the sensor 3, detects feature points of the work-piece shape from a plurality of waveform data W, and calculates a gripping position of the work-piece 7 on the basis of the feature points.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a system for instructing a gripping position of a workpiece to a robot arm, and an apparatus for picking the workpiece at a gripping position instructed by the system.

  Conventionally, when picking a workpiece placed on a work surface with a gripping claw of a robot arm, a method of photographing the workpiece to be grasped with a camera and detecting the position and orientation of the workpiece from the captured image is known. . For example, there is a method for detecting a position or posture of a workpiece by detecting a point that is a feature of the workpiece shape from a photographed image and performing pattern matching between the feature point on the image and the feature point on the workpiece model stored in advance. Are known.

  Patent Document 1 describes a technique in which a camera is used as the vision of an industrial robot and a feature point of a workpiece shape is obtained by processing a three-dimensional distance image captured by the camera. Patent Document 2 proposes a technique for adjusting the brightness and direction of illumination for illuminating a workpiece when a picking error occurs in a robot system using a stereo camera or a distance image sensor.

JP 2001-143073 A JP 2014-180708 A

  However, the method of obtaining feature points of a workpiece shape from image data of a camera using image processing or pattern matching is because the image data does not have depth information. It was not suitable for picking applications. In addition, since the camera image is easily affected by the lighting environment including ambient light, many types of workpieces with different surface properties (color and roughness), in-angle position, height, or depth under the same lighting environment. There was a problem that could not be picked.

  Therefore, an object of the present invention is to use a two-dimensional displacement sensor to accurately pick a variety of workpieces without being affected by the illumination environment and a system that can instruct picking data including depth information to the robot arm. To provide an apparatus.

In order to solve the above problems, the present invention provides the following characteristic means.
(1) A system for instructing a gripping position of a workpiece to a controller of a robot arm provided with a gripping portion movable in an XYZ three-dimensional space using a two-dimensional displacement sensor, the XZ plane from the two-dimensional displacement sensor An input unit that inputs a plurality of waveform data representing the workpiece shape in the Y direction, a calculation unit that calculates a gripping position of the workpiece by the gripping unit based on the plurality of waveform data, and a robot arm that receives control information including the workpiece gripping position An output unit that outputs to the controller of the computer, and the calculation unit includes a feature point detection unit that detects a feature point of the workpiece shape from the waveform data, and the workpiece gripping position in the XYZ three-dimensional space based on the detected feature point A gripping position indicating system characterized by obtaining

(2) The feature point detection unit obtains the change rate of the workpiece shape based on the waveform data, and detects the peak point farthest from the XY plane in the Z direction among the plurality of peak points of the change rate as the feature point of the workpiece shape A gripping position indicating system characterized by:

(3) A gripping position instruction system in which the feature point detection unit detects a gap portion of a work that the gripping unit of the robot arm can enter based on a change rate of the work shape.

(4) A gripping position instruction system in which the feature point detection unit obtains an angle at which the gripping unit of the robot arm enters the gap of the workpiece based on the waveform data.

(5) Using the gripping position instruction system described in any one of (1) to (4) above, one of a plurality of randomly stacked workpieces is picked by the gripping portion of the robot arm. A workpiece picking device characterized by that.

(6) The workpiece picking apparatus, wherein the plurality of workpieces include at least one irregular workpiece.

(7) A workpiece picking apparatus, wherein an XY plane on which a workpiece is placed and a two-dimensional displacement sensor are installed so as to be relatively movable.

  According to the gripping position instruction system of the present invention, there is an effect that picking data including depth information can be provided to the robot arm using the two-dimensional displacement sensor. In addition, according to the work picking apparatus of the present invention, there is an effect that various types of work can be accurately picked according to data provided from the gripping position instruction system without being affected by the illumination environment.

It is the schematic of the workpiece picking apparatus which shows one Embodiment of this invention. It is a block diagram which shows the system which instruct | indicates the holding position of a workpiece | work to a picking apparatus. It is a graph which illustrates the waveform data which the two-dimensional displacement sensor output. It is the graph which represented the change rate of the work shape corresponding to waveform data. It is a flowchart which shows operation | movement of a holding | grip position instruction | indication system.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. A workpiece picking apparatus 1 shown in FIG. 1 includes a robot arm having a two-dimensional displacement sensor 3 that scans a plurality of workpieces 7 arranged on an XY plane 4 and a gripping claw 6 that grips one of the plurality of workpieces 7. 5, a controller 8 that controls the movement of the robot arm 5, and a gripping position instruction system 2 that instructs the controller 8 on the gripping position of the workpiece 7 by the gripping claws 6 based on the waveform data W output from the two-dimensional displacement sensor 3. ing. The gripping claw 6 can move in an XYZ three-dimensional space. Depending on the type of the workpiece 7, the gripping claws 6 can be equipped so as to be rotatable around an axis extending in the Z direction.

  Although the work 7 is not limited to a specific shape and material, in this embodiment, the work 7 is an irregular work formed in a ring shape with a flexible material. The plurality of irregularly shaped workpieces 7 are randomly stacked on the XY plane 4 and arranged in an XYZ three-dimensional space in an aggregated state. The XY plane 4 is included in a part of a workpiece transfer surface formed by, for example, a conveyor, and moves in the Y direction with respect to the two-dimensional displacement sensor 3. The two-dimensional displacement sensor 3 irradiates a band-shaped laser beam extending in the X direction toward the XY plane 4, and outputs a large number of waveform data W representing the work shape in the XZ plane with relative movement with the XY plane 4. It is like that.

  As shown in FIG. 2, the gripping position instruction system 2 includes an input unit 11 that inputs waveform data W from the two-dimensional displacement sensor 3, and a gripping position of the workpiece 7 by the gripping claws 6 based on the input waveform data W. 1 and an output unit 15 that outputs control information including the calculated gripping position to the controller 8 of the robot arm 5, each unit being mounted on the computer 10 shown in FIG. The arithmetic unit 12 is provided with a feature point detection unit 13 that detects feature points of the workpiece shape from the waveform data W on the XZ plane.

  Next, the process in the calculating part 12 is demonstrated more concretely according to FIG. 3 and FIG. As shown in FIG. 3 (a), the two-dimensional displacement sensor 3 is a band-shaped assembly of workpieces 7 randomly stacked on the workpiece conveyance surface in accordance with relative movement with the XY plane 4 (see FIG. 1). As shown in FIG. 3B, the waveform data (profile data) W1, 2,..., N corresponding to a large number of scanning lines L1, 2,. And the acquired waveform data W is output to the computer 10.

  In the computer 10, the input unit 11 inputs the waveform data W from the two-dimensional displacement sensor 3, and the feature point detection unit 13 of the calculation unit 12 detects the feature point of the workpiece shape from the waveform data W. FIG. 4 shows how to obtain feature points by taking the waveform data W2 of FIG. 3B as an example. The feature point detector 13 obtains the rate of change in height (ΔZ / ΔX) in the Z direction for each minute distance in the X direction (for example, 0.005 to 0.01 mm) in the waveform data W2, and plots the rate of change. A change rate graph W2 ′ is generated. And while detecting the peak point 23max farthest in the Z direction from the XY plane 4 among the plurality of peak points 23 in the change rate graph W2 ′, based on the change rate of the work shape, The grip portion 6 of the robot arm 5 detects a gap that can enter the workpiece assembly.

  In this way, the position at which the workpiece 7 can be gripped by the gripping claws 6 can be uniquely determined in the workpiece assembly, so that it is not necessary to grasp the position, shape, or posture of each workpiece. 21, the gripping position of the workpiece 7 can be easily calculated. In calculating the gripping position, the calculation unit 12 calculates the XYZ coordinate value of the feature point 21, that is, the X and Z coordinate values in the waveform data and the Y coordinate value of the scanning line L2 corresponding to the waveform data in an XYZ three-dimensional manner. It is obtained as a gripping position of the work 7 in the space, and is provided from the output unit 15 to the controller 8 of the robot arm 5. Further, by examining the front-rear relationship of the feature points, the orientation of the gripping claws 6 can be analyzed, and the angle at which the gripping portion 6 of the robot arm 5 enters the gap portion of the workpiece 7 can be obtained based on the waveform data. . In FIG. 4, two peak points 23 max are shown at the same height, but one of them is selected as the feature point 21.

  Next, the operation of the gripping position instruction system 2 will be described with reference to FIG. When the workpiece picking device 1 is activated, the gripping position instruction system 2 starts an operation for instructing the controller of the robot arm 5 about the gripping position of the workpiece 7. First, the gripping position instruction system 2 inputs waveform data W in the XZ plane from the two-dimensional displacement sensor 3 (S51 in FIG. 5). Next, feature points of the workpiece shape are detected based on the input waveform data W (S52).

  Subsequently, the gripping position instruction system 2 detects a peak point farthest in the Z direction from the XY plane 4 among the plurality of feature points as an optimum feature point, and the gripping claw 6 can enter near the feature point. A void is detected (S53). Then, the orientation of the gripping claw 6 is obtained from the front-rear relationship of the feature points, the approach angle of the gripping claw 6 into the gap is calculated from the waveform data (S54), and control information including the calculation result is output to the controller 8 ( S55). Note that the processing shown in FIG. 5 is merely an example, and the processing order can be changed as appropriate.

  Therefore, according to the gripping position instruction system 2 of this embodiment, the control information is output based on the waveform data W output from the two-dimensional displacement sensor 3, so that the depth information is different from the conventional image processing and pattern matching. Including picking data can be provided to the controller 8 of the robot arm 5. For this reason, the workpiece picking apparatus 1 can pick the workpieces 7 one by one in order from the moving workpiece assembly randomly stacked on the XY plane 4 by the gripping claws 6 of the robot arm 5.

  Further, according to the gripping position instruction system 2 of this embodiment, since the point farthest from the XY plane 4 is detected as the optimum feature point of the workpiece shape, the gripping claw 6 can be controlled regardless of the shape of the workpiece 7 itself. It is possible to easily calculate the position, orientation, and approach angle. As a result, for example, it is possible to pick an irregular workpiece that has been difficult by pattern matching, such as a flexible packing, without mistakes by the gripping claws 6 of the robot arm 5, and the irregular workpiece that has been relied on by a conventional human hand. There is also an effect that the picking work of the work can be automated.

  The present invention is not limited to the above embodiment, and the rotation angle around the Z-axis of the gripping claws 6 is controlled based on the waveform data, or a vacuum or magnetic attraction head is used instead of the gripping claws 6. For example, the configuration of each unit can be arbitrarily changed without departing from the spirit of the present invention.

DESCRIPTION OF SYMBOLS 1 Work picking apparatus 2 Gripping position instruction system 3 Two-dimensional displacement sensor 4 XY plane 5 Robot arm 6 Gripping claw 7 Work 8 Controller 10 Computer 11 Input unit 12 Calculation unit 13 Feature point detection unit 15 Output unit 21 Feature point 23 Change rate Peak point

Claims (7)

A system for instructing a gripping position of a workpiece to a controller of a robot arm having a gripping portion movable in an XYZ three-dimensional space using a two-dimensional displacement sensor,
An input unit that inputs a plurality of waveform data representing a workpiece shape in the XZ plane from the two-dimensional displacement sensor in the Y direction; a calculation unit that calculates a gripping position of the workpiece by the gripping unit based on the plurality of waveform data; An output unit that outputs control information including the gripping position to the controller;
The calculation unit includes a feature point detection unit that detects a feature point of the workpiece shape from the waveform data, and obtains a gripping position of the workpiece in the XYZ three-dimensional space based on the feature point. Position indication system.   The feature point detection unit obtains a rate of change of the workpiece shape based on the waveform data, and among the plurality of peak points of the rate of change, a peak point farthest from the XY plane in the Z direction is a feature of the workpiece shape. 2. The grip position indicating system according to claim 1, wherein the grip position indicating system is detected as a point.   The gripping position instruction system according to claim 1, wherein the feature point detection unit detects a gap portion of the work that the gripping unit of the robot arm can enter based on a change rate of the work shape.   4. The gripping position instruction system according to claim 3, wherein the feature point detection unit obtains an angle at which the gripping unit of the robot arm enters the gap of the workpiece based on the waveform data.   Using the grip position indicating system according to any one of claims 1 to 4, one of a plurality of workpieces randomly stacked on the XY plane is picked by a grip portion of the robot arm. A workpiece picking device characterized by that.   6. The workpiece picking apparatus according to claim 5, wherein the plurality of workpieces include at least one irregular workpiece.   7. The workpiece picking apparatus according to claim 5, wherein the XY plane on which the workpiece is placed and the two-dimensional displacement sensor are installed so as to be relatively movable.

Images