JPH06155349A - Work identification device for robot - Google Patents

Abstract

PURPOSE:To reduce the number of stops of a robot even if a camera fails in a robot system using a visual sensor such as camera, etc. CONSTITUTION:A fixed camera 5 is fixed over a work 3, a hand camera 15 is mounted on a hand 2 of a robot 1, and an image signal from the fixed camera 5 is sent to an image processing device 6 to identify the work 3. If mis- identification or damage of the work 3 by the fixed camera 5 is detected, signal transmission from the fixed camera 5 is switched over to that from the hand camera 15, and the work 3 is identified again. Instead of the hand camera 15, a beam sensor may be used, and also the hand camera 15 and beam sensor may be installed on the hand 2.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a device for a robot to recognize a work by a visual sensor such as a camera.

[0002]

2. Description of the Related Art FIG. 7 is a block diagram showing a conventional robot work recognition device disclosed in, for example, Japanese Patent Laid-Open No. 3-95025. In the figure, (1) is a robot, (2) is the hand of the robot (1), (3) is a work (a cube box or similar object) stacked on a pallet (4), and (5) is a work. A fixed camera installed above (3), (6) an image processing device connected to the fixed camera (5), (7) a robot (1)
Robot controller connected to, (8) is a ground controller for controlling the traveling of the pallet (4), (9) is each of the above devices (6) to (8)
The integrated control device is connected to the computer, and (10) is a host computer connected to the integrated control device (9).

A conventional work recognition device for a robot is constructed as described above, and a self-propelled carriage (not shown) in which works (3) are stacked on a pallet (4) is controlled by a ground controller (8). Are conveyed to a predetermined position. Robot (1)
Is controlled by the robot controller (7), and the transferred works (3) are transferred to the conveyor (not shown) or the like by the hand (2). Where the fixed camera (5) is the work
(3) is imaged and its image signal is sent to the image processing device (6). The image processing device (6) works from the input image signal.
The array pattern of (3) is recognized and an array pattern signal is output to the overall control device (9). Then, the integrated control device (9) is a robot control device based on the input array pattern signal.
A command will be given to (7).

[0004]

In the conventional work recognition device for robots as described above, the work is fixed by the fixed camera (5).
Since the image of (3) is taken and the robot (1) is controlled based on this, there is a problem that the fixed camera (5) erroneously recognizes the work (3). Further, if the fixed camera (5) is damaged, the system will go down, the machine will stop, and it will be necessary to manually restore it.

The present invention has been made in order to solve the above-mentioned problems, and it is possible to obtain correct image recognition even if the fixed camera is erroneously recognized, and to stop the robot even if the fixed camera is damaged. An object of the present invention is to provide a work recognition device for a robot that can be avoided.

[0006]

A work recognition device for a robot according to a first invention of the present invention has a fixed camera fixed above the work and a hand camera mounted on the robot. Whether the image processing device that recognizes the work imaged by the camera and the fixed camera misrecognize the work,
Alternatively, it is provided with a failure detecting means for detecting the breakage and a switching means for switching the fixed camera to the hand camera to recognize the work again when the failure detecting means operates.

A work recognition device for a robot according to a second aspect of the present invention includes a fixed camera fixed above the work,
A position detection sensor mounted on the robot, an image processing device for recognizing the work imaged by the fixed camera, a failure detection unit for erroneously recognizing the work by the fixed camera, or a damage detection unit, and the failure detection unit. Is operated, the switching means switches from the fixed camera to the position detection sensor to recognize the work again.

A work recognition device for a robot according to a third aspect of the present invention includes a hand camera and a position detection sensor mounted on the robot, an image processing device for recognizing the work imaged by the hand camera, and a hand camera for detecting the work. It is provided with failure detection means for erroneously recognizing or for detecting damage, and switching means for switching from the hand camera to the position detection sensor to recognize the work again when the failure detection means operates.

[0009]

According to the first aspect of the present invention, when a failure of the fixed camera is detected, the hand camera is switched to recognize the work. Therefore, the hand camera can continue the image recognition without stopping the system. Complete.

Further, in the second invention, when a failure of the fixed camera is detected, the position detection sensor is switched to, and in the third invention, when the failure of the hand camera is detected, the position detection sensor is switched to. Since the workpieces are individually recognized, the position detection sensor can continue to recognize the workpieces, and the system does not have to be stopped. In addition, the camera and the position detection sensor are completely different in software and hardware.

[0011]

【Example】

Example 1. 1 to 3 are views showing an embodiment of the first invention of the present invention, FIG. 1 is a block diagram, FIG. 2 is an image image explanatory view, and FIG. 3 is a recognition operation flowchart, which is the same as the conventional device. The parts are denoted by the same reference numerals (the same applies to the following embodiments).

In FIGS. 1 and 2, (15) is a robot
The hand camera attached to the hand (2) of (1), (16) is a television (hereinafter referred to as TV) monitor, and A is a fixed camera.
(5) is an image, B is an image by the hand camera (15), and C1 and C2 are images of the work (3) to be processed.

Next, the operation of this embodiment will be described with reference to FIG. In step (21), the fixed camera (5) recognizes the work (3). That is, work with a fixed camera (5)
When (3) is imaged, the upper surface of the entire work (3) is within the field of view, so that an image C1 as shown in image A of FIG. 2 is obtained. The image signal corresponding to the image C1 is an image processing device.
It is sent to (6) and processed, and the TV monitor (1
Displayed in 6).

In step (22) (corresponding to failure detecting means), it is judged whether the image C1 is correct or not based on the image signal. If the result is positive, the process jumps to step (26), and the hand (2) is operated to grip the work (3) (the description of the subsequent operation is the same as that of the conventional device, and therefore omitted). If the result is negative, the process proceeds to step (23) (corresponding to the switching means) to switch to the recognition of the work (3) by the hand camera (15).
At this time, a part of the work (3) is placed in the field of view, and the work (3) is locally imaged. Here, the image C2 as shown in the image B of FIG. 2 is obtained.

At step (24), it is judged whether the image C2 is right or wrong. If the result is positive, the process jumps to step (26) and the work (3) is gripped by the hand (2). If the result is negative, the process proceeds to step (21) to execute unrecognizable processing as erroneous recognition, and stop the robot (1), for example.

Example 2. 4 and 5 show a second embodiment of the present invention.
FIG. 4 is a block diagram and FIG. 5 is an operation explanatory view of the beam sensor. Note that FIG. 3 is also used in this embodiment.

In this embodiment, the hand camera (15) shown in FIG. 1 is replaced with a beam sensor (18). The beam sensor (18) detects the object by projecting the projection beam (18a) and detecting the reflected light thereof, and constitutes a position detection sensor. That is, as shown in FIG. 5, with the beam sensor (18) projecting the projection beam (18a), the hand (2) is moved in the scanning direction X for scanning. At this time, if the projection beam (18a) is blocked by the work (3),
The beam sensor (18) signal goes from off to on.

The robot controller (7) reads the coordinate value of the robot (1) at the changed point of this signal,
You will know the position of one side of the work (3). Furthermore, the position of one side of the detected work (3) and the side in the right angle direction (see FIG.
Similarly, the direction Y) which is perpendicular to the scanning direction X of is detected.
The position of the work (3) is detected from the detected positions of the two sides.

The operation of this embodiment is similar to that of FIG. However, in step (23), the beam sensor (18) recognizes the work (3). As is clear from FIG. 4, the signal processing systems of the fixed camera (5) and the beam sensor (18) are independent of each other.

Example 3. FIG. 6 is a block diagram showing an embodiment of the third invention of the present invention. 3 and 5 are also used in this embodiment.

In this embodiment, instead of using the fixed camera (5) of FIG. 4, a hand camera (15) is attached to the hand (2). Also in this embodiment, as is apparent from FIG. 6, the signal processing systems of the hand camera (15) and the beam sensor (18) are independent of each other. The operation of this embodiment is similar to that of FIG. However, the work (3) is recognized by the hand camera (15) in step (21), and the work (3) is recognized by the beam sensor (18) in step (23).

Example 4. In each of the above embodiments, the fixed camera
Although the combination of two of (5), the hand camera (15) and the beam sensor (18) is shown, a more reliable system can be constructed by combining these three.

Example 5. Although the beam sensor (18) of the second and third embodiments has been described as a sensor that applies light, a sensor that applies ultrasonic waves, magnetism, etc. is also useful. Further, even a contact type sensor can be used as long as the presence or absence of an object can be clearly detected.

[0024]

As described above, according to the first aspect of the present invention, when the failure of the fixed camera is detected, the hand camera is switched to recognize the work, so that the hand camera can continue the image recognition. , The system does not need to be stopped, which has the effect of increasing reliability and improving productivity.

Further, in the second invention, when the failure of the fixed camera is detected, the position detection sensor is switched to, and in the third invention, when the failure of the hand camera is detected, the position detection sensor is switched to, Since each work is recognized, the work can be continuously recognized by the position detection sensor, and the system does not have to be stopped. In addition, the processing system of the camera and the position detection sensor are completely independent, which has the effect of further improving reliability and productivity.

[Brief description of drawings]

FIG. 1 is a block diagram showing a first embodiment of the present invention.

FIG. 2 is an explanatory diagram of an image image according to FIG.

FIG. 3 is a recognition operation flowchart of FIG. 1.

FIG. 4 is a block diagram showing a second embodiment of the present invention.

FIG. 5 is an operation explanatory view of the beam sensor according to the second and third embodiments of the present invention.

FIG. 6 is a block diagram showing Embodiment 3 of the present invention.

FIG. 7 is a block diagram showing a conventional work recognition device for a robot.

[Explanation of symbols]

 1 Robot 2 Hand 3 Work 5 Fixed Camera 6 Image Processing Device 7 Robot Controller 15 Hand Camera 18 Beam Sensor

Claims (3)

[Claims] 1. An image for recognizing the work, which has a fixed camera fixed above the work and picks up the work, and a hand camera attached to a robot hand to pick up the work. A processing device, a failure detection unit that detects that the fixed camera erroneously recognizes the work piece or that the fixed camera is damaged, and when the failure detection unit operates, the fixed camera is switched to the hand camera and again. A work recognition device for a robot, comprising: a switching means for recognizing the work. 2. A fixed camera fixed above a work for picking up the work, a position detection sensor mounted on a hand of a robot for detecting the position of the work, and an image for recognizing the work picked up by the fixed camera. A processing device,
Failure detection means for detecting that the fixed camera erroneously recognizes the work or damage to the position detection sensor, and when the failure detection means operates, the fixed camera is switched to the position detection sensor and the work is again detected. A work recognition device for a robot, comprising: a switching means for recognizing the work. 3. A hand camera mounted on a hand of a robot for imaging a work, a position detection sensor mounted on the hand for detecting the position of the work, and an image processing device for recognizing the work imaged by the hand camera. And a failure detecting means for detecting that the hand camera is damaged, and a switching means for switching the hand camera to the position detecting sensor to recognize the work again when the failure detecting means operates. Work recognition device.