JP2010162638A - Robot hand - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To accurately select an object and hold the object. <P>SOLUTION: A robot hand includes: a support part 40; first and second holding parts 110 oppositely arranged at the support part 40 and holding the object 100; first and second light sources 90 oppositely arranged at the first and second holding parts 110 respectively and irradiating light to the object 100; an imaging device 50 being arranged at a support part 40, being arranged at an area sandwiched by planes including the first and second holding parts 110, and obtaining the image of the object 100 by imaging the object 100 irradiated by light; and an image processing device connected with the imaging device 50 and processing the image obtained by the imaging device 50. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a robot hand, and more particularly to a robot hand that can accurately select an object from scattered objects and can grip the object.

  With the progress of automation, rational work using industrial robots is indispensable. Among them, techniques for gripping, for example, box-shaped objects with a robot hand have been actively studied.

  In a conventional robot hand, a control method and apparatus have been invented in which an optical sliding detector is provided on a finger and the gripping force of the finger is automatically determined by the sliding detector (for example, a patent) Reference 1).

JP 2005-153114 A

  However, the robot hand described in Patent Document 1 has a problem that it is difficult to appropriately select an object from among scattered objects even though the object can be gripped.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a robot hand that can accurately select an object and can grip the object.

In order to achieve the above object, a robot hand according to the present invention is provided with a support portion, first and second gripping portions that are provided on the support portion and are opposed to grip an object, and the first and second First and second light sources facing each other for irradiating the object with light,
An image of the object is obtained by imaging the object that is provided in the support and provided in an area sandwiched between planes including the first and second gripping parts and irradiated with the light. The robot hand includes an imaging device to be acquired and an image processing device that is connected to the imaging device and processes the image acquired by the imaging device.

  According to the present invention, it is possible to accurately select an object from among the scattered objects and hold the object.

The figure which shows 1st Embodiment of the robot hand concerning this invention. The figure of the robot hand concerning a 1st embodiment. The figure which shows the drive mechanism of the joint part concerning 1st Embodiment. The figure when irradiating an object with light by ring illumination. The block diagram when installing a light source in the left side of the object concerning 1st Embodiment, and installing the imaging device above the object. The block diagram when installing a light source in the right side of the object concerning 1st Embodiment, and installing the imaging device above the object. FIG. 6 is a superimposed diagram of edge images obtained by imaging in FIGS. 4 and 5. FIG. 5 is a diagram for explaining an object imaging method according to the first embodiment. The figure of the robot hand concerning a 2nd embodiment. The figure of the robot hand concerning a 2nd embodiment. The figure of the robot hand concerning a 2nd embodiment. The flowchart of rotation, irradiation, and imaging of the robot hand concerning 2nd Embodiment. The figure explaining the differentiation process of the image concerning 2nd Embodiment. The figure explaining the differentiation process of the image concerning 2nd Embodiment. The figure explaining the differentiation process of the image concerning 2nd Embodiment. The figure explaining the differentiation process of the image concerning 2nd Embodiment. Explanatory drawing about the photometric at the time of recognition of the sphere concerning 2nd Embodiment. Explanatory drawing about the photometric at the time of recognition of the sphere concerning 2nd Embodiment. The figure which shows the projection angle of an object. The figure which shows 3rd Embodiment of the robot hand to which this invention is applied. The figure which shows the 1st modification of 3rd Embodiment. The figure which shows the 2nd modification of 3rd Embodiment. The figure which shows 4th Embodiment of the robot hand concerning this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  In the drawings to be described below, the same reference numerals indicate the same parts, and duplicate descriptions are omitted.

(First embodiment)
FIG. 1A is a diagram showing a first embodiment of a robot hand according to the present invention.

  In the robot hand shown in this embodiment, the arm portion 10 and the wrist portion 30 are connected via the joint portion 20, and the rod-like support portion 40 is provided on the wrist portion 30.

  The support portion 40 is provided with two gripping portions 110, and the two gripping portions 110 are provided in a positional relationship in which they face each other. In addition, the support unit 40 is provided with an imaging device 50 for detecting the object 100. The arm part 10 and the wrist part 30 can move flexibly so that the grip part 110 can be positioned in any direction by the joint part 20.

  Further, the gripping part 110 includes a slide guide 60, a spring 70, and a rod-like finger part 80, and a light source 90 for irradiating the object 100 is provided on the tip part.

  Also, there may be a plurality of gripping portions 110. For example, the gripping portion 110 may have four configurations as shown in FIG. FIG. 1B shows a view of the robot hand as viewed from above. For the sake of simplicity, components other than the support portion 40, the light source 90, and the grip portion 110 are omitted. That is, the light source 90 provided on the other gripping part 110 is provided on the opposite side of the light source 90 provided on one gripping part 110, and light is emitted from each light source 90 toward the object 100. I hope you can.

  FIG. 2 is a diagram illustrating an example of a drive mechanism of the joint unit 20. As shown in FIG. 2, the driving force of the motor 11 provided in the arm portion 10 is decelerated by the speed reducer 12, and the driving force is transmitted to the bevel gear 21 provided in the joint portion 20 and rotated. The wrist part 30 can be moved.

  The support part 40 in FIG. 1A is fixed to the wrist part 30 and includes two gripping parts 110 for gripping the object 100. The mechanism that opens and closes the gripping part 110 (the mechanism that moves in the horizontal direction) may include, for example, a laterally fed ball screw and a stepping motor in the support part 40, and may be opened and closed by driving the stepping motor.

  Further, the imaging device 50 is attached to the support portion 40 in a region sandwiched between planes including two opposing gripping portions 110. The imaging device 50 is preferably provided, for example, between the two gripping units 110 and in the center of the support unit 40, but may be provided anywhere on the support unit 40 as long as the object 100 can be imaged. . Each gripping part 110 includes a slide guide 60, a spring 70, a finger part 80, and the like in order to grip the object 100. The slide guide 60 is attached to the support portion 40 and houses a part of the spring 70 and the finger portion 80 therein. Then, it acts as a guide for the moving direction (vertical direction) when sliding the finger 80. Further, when the object 100 is gripped, an operation of gripping the object 100 using the finger unit 80 is performed.

  The spring 70 is provided between the support part 40 and the finger part 80 in the slide guide 60. Then, it acts to absorb an impact when the finger 80 comes into contact with the object 100.

  Further, in order to grasp the amount of displacement of the finger 80, a displacement sensor may be provided in parallel with the spring 70 to control the position of the finger 80. In addition, the height of the wrist part 30 is obtained from the angle of the joint part 20 of the arm part 10 by the ON / OFF switch, and the finger part 80 is controlled by monitoring this ON / OFF switch. Also good.

  Furthermore, a light source 90 for illuminating the object 100 is provided at the tip of each finger unit 80. If the light source 90 can illuminate the object 100, at least one of the finger portions 80 is provided for the support unit 40 to freely move and irradiate the object 100 from various directions. Just do it. Examples of the light source 90 include a light emitting diode.

  Next, the operation principle of the robot hand will be described.

  As shown in FIG. 1A, the robot hand according to the present embodiment illuminates the object 100 with the light source 90 provided on each finger unit 80 and recognizes the object 100, and is provided on the support unit 40 of the robot hand. The object 100 is captured by the imaging device 50 and the image is stored.

  At this time, as shown in FIG. 3, for example, as shown in FIG. 3, the structure that illuminates the object 100 from the upper side of the object 100 by using, for example, the ring illumination 120 so as not to make a shadow as a whole. Met.

  However, in this method, as shown in the right diagram of FIG. 3, the image obtained by the imaging device 50 has a sharp edge of the object 100 in the image obtained by the imaging device 50, but the image 100 is equally viewed from above. Since the light is irradiated, the inner edge becomes an unclear image, and the object 100 may not be identified. In other words, this method is not suitable for extracting all the inner edge portions of the object 100 because it does not make a shadow as a whole.

  In the present embodiment, the object is irradiated from two steps. That is, in the first step, as shown in FIG. 4, the object 100 is irradiated with light from a light source 90 installed on the left side of the object 100, for example, and the object 100 is imaged by the imaging device 50 installed above the object 100. . As shown in the diagram on the right side of FIG. 4, the image obtained by this is a sharp image at the inner edge 100 a on the side close to the light source 90, and a clear image at the inner edge 100 b on the side away from the light source 90. Become. This is because light does not easily hit the right side (the side far from the light source 90) of the object 100, and there is a difference in brightness between the upper surface portion and the right side surface portion of the object 100. In general, in image processing, edge extraction is performed by luminance differentiation, and an edge at a place where the contrast is clear is easy to extract.

  Next, in the second step, the object 100 is irradiated with light from the light source 90 installed on the right side of the object 100, and the object 100 is imaged by the imaging device 50 installed above the object 100. In this case, as shown in the diagram on the right side of FIG. 5, a sharp image can be obtained with the left inner edge 100c, but a sharp image can be obtained with the right inner edge 100d.

  Then, as shown in FIG. 6 (a), by superimposing the image obtained in the first step of FIG. 4 and the image obtained in the second step of FIG. Therefore, a clear image can be obtained.

In other words, a plurality of images of the object 100 are captured using the imaging device 50 from above with respect to the emission direction of the light emitted from the light source 90 emitted to the object 100, and edge images of these images are extracted. What is necessary is just to perform the process which synthesize | combines an edge image. The imaging device 50 can obtain a clear image by imaging the object 100 from above with respect to the emission direction of the light from the light source 90 emitted to the object 100. However, in this case, in order to obtain a clearer image, the elevation angle and the depression angle are about 45 ° or more from the light emission direction from the light source 90 emitted to the object 100 from the light emission direction. If an imaging device is provided, a clear image can be extracted. This state is shown in FIG. FIG. 6B shows a region 112 in which a clear image can be extracted with respect to the emission direction 111 of light from the light source 90 (the arrow shown in FIG. 6B).
(A region where the elevation angle and depression angle with respect to the emission direction 111 are 45 ° or more). If the imaging device 50 is provided in this area 112, a clear image can be extracted.

(Second Embodiment)
7 to 9, the object 100 is irradiated with light emitted from the light sources 90 of the two gripping units 110 provided on the support unit 40 while rotating the support unit 40 of the arm unit 10 to support the upper side. A state in which an image is picked up from above with respect to the direction of the light emitted from each light source 90 by the imaging device 50 provided in the unit 40 is shown. For simplicity, details of the arm portion 10 to the wrist portion 30 and the gripping portion 110 are omitted.

  8 shows a view in which the support portion 40 is rotated by 45 ° from the position of FIG. 7, and FIG. 9 shows a view in which the support portion 40 is rotated by 90 ° from the position of FIG. In addition, the figure seen from the upper direction is shown in the figure on the right side of FIGS.

  FIG. 10 shows an imaging process of the object 100 according to the present embodiment and a flowchart for acquiring an image.

  First, as shown in FIG. 7, when an image is acquired using the imaging device 50, the arm unit 10 of the robot hand approaches the target object 100 (step S100 in FIG. 10). Next, light is emitted from the first lateral direction of the object 100 by the light source 90 provided in the gripping unit 110 of the arm unit 10, and a captured image of the object 100 is acquired by the imaging device 50. An edge is extracted by performing an image differentiation process on the acquired image by an image processing device (not shown) (step S110 in FIG. 10). Let this edge image be A.

  The image processing apparatus may be built in the imaging apparatus 50 or may be externally connected to the robot hand. Moreover, if the imaging device 50 is provided in the plane where the two gripping portions 110 of the support portion 40 face each other and is provided in the vicinity of the center, the step of performing the coordinate conversion processing of the captured image can be omitted. preferable. However, even if it is provided at other positions, for example, an in-plane region facing the plane including the two gripping portions 110, or other positions, the imaging apparatus 50 may be aligned by coordinate conversion processing. .

  For example, as shown in FIG. 11, when the white window 300 is displayed at the center of the gray background 200, the image differentiation processing is performed by setting the luminance on the line segment II ′ to the z axis (in the direction perpendicular to the paper surface). ), The horizontal coordinate of the screen is x, and the graph is written as shown in FIG. This is regarded as a function z = f (x). The function f changes arbitrarily depending on the pattern.

  Then, when z ′ = f ′ (x) obtained by differentiating the function z = f (x) with respect to x is drawn as shown in FIG. 13, an image as shown in FIG. 14 is obtained as a screen. By this differentiation processing, the edge portion of the image can be extracted.

  In addition to the method of extracting an edge by differentiating an image, it is possible to grasp the shape of an object using a photometric method.

  For example, when a sphere 120 as shown in FIG. 15 is recognized, the shape of the sphere 120 becomes unclear only with a fixed light source and a fixed imaging device. Accordingly, by moving the light source 90 from the position state of the light source 90 in FIG. 15 to the position in FIG. 16, the shine position of the sphere 120 (the position where the reflected light is strong) changes, and it is easily recognized that it is a sphere. be able to. This method is not limited to a sphere, but can be applied to an arbitrary shape such as an ellipsoid or a gourd. In other words, this method is based on a technique in which the shape of an object can be recognized by applying light to the object from various directions in the dark.

  Next, as shown in FIG. 8, the grip 110 is rotated 45 ° from the state of FIG. 7, and light is irradiated from the second lateral direction in the same manner as described above. 100 images are acquired (step S120 in FIG. 10). Let this edge image be B.

  Further, as shown in FIG. 9, the gripping part 110 is rotated 90 ° from the state of FIG. 7, and light is irradiated from the third lateral direction in the same manner as described above, and the object 100 from the upper imaging device 50 is irradiated. Is acquired (step S130 in FIG. 10). Let this image be C. Further, an image obtained when the gripping part 110 is rotated to the fourth angle is acquired by the same method as described above (step S140 in FIG. 10), and the image is set as D.

  Then, these four images A, B, C, and D are superimposed to form an image that captures as many edges of the object 100 as possible (step S150 in FIG. 10). The number of images to be picked up depends on the shape of the object to be picked up, but may be at least two in 90 ° increments, for example.

  Next, the model image of the object 100 captured in advance is matched with the actually acquired image (step S160 in FIG. 10), and the position and shape of the object 100 are calculated by recognizing the object 100 (in FIG. 10). Step S170).

  For example, as shown in FIG. 17, if the projection angles α, β, γ on the screen of the object are known, the three-dimensional shape of the object can be determined without depending on the distance between the object and the camera. Can be calculated. This is because the angle is invariant with distance. Once the shape is obtained, the distance between the object and the imaging device can be calculated by comparing the length of the object side with the object model.

  In this way, the object 100 is irradiated with light from various angles, and an image is acquired by imaging the object 100 from above with respect to the emission direction of the light emitted from the light source 90, and the images are acquired. By superimposing these images, the shape of the object 100 can be accurately recognized.

(Third embodiment)
FIG. 18 shows a third embodiment of a robot hand to which the object detection method of the present invention is applied.

  The difference from the first embodiment is that the finger unit 80 used in the robot hand of the first embodiment is a light guide unit 130.

  The light guide unit 130 is made of a transparent material, and the tip thereof is processed so that light emitted from the light source 90 is totally reflected by the tip of the light guide unit 130 and is irradiated onto the object 100. As the transparent material, for example, a material such as acrylic or glass can be used, and it is necessary to process the tip of the light guide unit 130 according to the material. For example, when acrylic is used, the refractive index is 1.49, the total reflection angle is 42.1 ° according to Snell's law, and the tip of the light guide unit 130 may be cut at least 42.1 °. In addition, when the above transparent material is used, the light guide unit 130 can be processed to be elongated because the processing is easy. By processing the light guide section 130 to be elongated, the hand becomes lighter, so the motor of the robot arm can be miniaturized, and it is possible to expect agile movement and improved work efficiency in a narrow space.

  By using the finger part 80 of FIG. 1A as the light guide part 130, the object 100 can be selected with higher accuracy by preventing the field of view of the imaging device 50 from being transparent because the light guide part 130 is transparent. There is a feature that can be done. At this time, the light source 90 may be provided at a position where an impact caused by the light guide 130 holding the object 100 is not transmitted to the light source 90.

  Further, as a first modification of the present embodiment, as shown in FIG. 19, the outer peripheral part of the light guide part 130 is covered with a covering part 140. Thereby, the light guide part 130 can be protected without being damaged. Examples of the covering portion 140 include a metal case and a metal coat. At this time, it is necessary to open the part where the light of the covering part 140 covering the tip of the light guide part 130 is emitted.

  Furthermore, as a third modification of the present embodiment, a transparent contact sensor 150 may be provided at the tip of the light guide unit 130 as shown in FIG. As the contact sensor 150, for example, a resistance contact sensor, a capacitance contact sensor, a piezoelectric element, or the like may be used. Thereby, contact with the object 100 can be confirmed without blocking light transmission.

(Fourth embodiment)
FIG. 21 shows a fourth embodiment of a robot hand to which the present invention is applied.

  The robot hand of the present embodiment is provided with a hand 160 for gripping the object 100 in a region sandwiched between a plurality of planes including the two gripping portions 110 installed on the support portion 40. The hand 160 is attached to the support unit 40.

  Also in this embodiment, the object 100 is detected by the method described above. In this case, the grip part 110 is fixed to the support part 40, and sorts the object 100 with high accuracy. For example, when there are a plurality of objects 100, the object is selected by the holding unit 100, and the selected object 100 is held by the hand 160.

  Further, the grip portion 110 does not need to be fixed to the support portion 40, and the object 100 is gripped by the grip portion 110 or the hand 160 depending on the situation. For example, when there are a plurality of objects 100 and there is a small object 100 that cannot be grasped by the grasping unit 110, the object 100 is collected to some extent by the grasping unit 110, and a hand 160 is used to pick up the plurality of objects 100. A desired object 100 is gripped from the inside. A plurality of hands 160 may be provided. The hand 160 may be a robot hand used in the present invention.

  The present invention is not limited to the above-described embodiment, and can be appropriately changed in design without departing from the gist of the invention. Also, the above-described first to fourth embodiments and modifications thereof are possible. May be combined as appropriate.

DESCRIPTION OF SYMBOLS 10 ... Arm part 11 ... Motor 12 ... Reduction gear 20 ... Joint part 21 ... Bevel gear 30 ... Wrist part 40 ... Support part 50 ... Imaging device 60 ... Slide guide 70 ... Spring 80 ... Finger part 90 ... Light source 100 ... Object 100a ... inner edge 100b ... inner edge 100c ... inner edge 100d ... inner edge 110 ... gripping part 111 ... emitting direction 112 ... region 120 ... sphere 130 ... light guide part 140 ... clothing part 150 ... contact sensor 160 ... hand

Claims (5)

A support part;
Opposing first and second gripping portions provided on the support portion for gripping an object;
First and second light sources provided on the first and second gripping units, respectively, for irradiating the object with light;
An imaging device that is provided in the support unit and captures the object irradiated with the light to obtain an image of the object;
A robot hand, comprising: an image processing device connected to the imaging device and processing the image acquired by the imaging device. A rotatable support, and
Opposing first and second gripping portions provided on the support portion for gripping an object;
A light source for irradiating the object with light at a plurality of positions provided on at least one of the opposing surfaces of the first and second gripping units and rotated by the support unit;
An imaging device that is provided in the support unit and captures the object irradiated with the light to obtain an image of the object;
A robot hand, comprising: an image processing device connected to the imaging device and processing the image acquired by the imaging device.   The first and second gripping portions are formed of a transparent light guide portion, and the tip of the light guide portion is processed at an angle that totally reflects light emitted from the first and second light sources. The robot hand according to claim 1, wherein the robot hand is a robot hand.   The robot hand according to claim 1 or 2, wherein the first and second gripping portions are provided with contact sensors on the inner surfaces of the first and second gripping portions facing each other.   3. The robot hand according to claim 1, wherein the support portion includes a hand in a region sandwiched between planes including the first and second gripping portions.

Images