JP2010276447A - Position measuring apparatus, position measuring method and robot system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a position measuring apparatus capable of precisely measuring a position of a measurement object even when an image of the measurement object photographed is blurry. <P>SOLUTION: The position measuring apparatus 1 includes a stage 72 on which the measurement object W is set and which has a marker 90 for emitting a plurality of refection light portions, a light irradiation section 48 for irradiating the marker 90 with light, a plurality of imaging devices 20, 21 for photographing the measurement object W and the marker 90 in different directions of sight line, a point spread function computing section 36 which extracts a marker image 100 of the marker 90 from a photographed image B taken by the imaging devices 20, 21 and obtains a point spread function from the marker image 100, an image generating section 37 for generating a transformed image G by applying an image transformation using the point spread function to the photographed image B, and a position calculating section 39 for calculating the position of the measurement object W on the basis of the transformed image G generated. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a position measurement device, a position measurement method, and a robot system including the position measurement device that measure the position of an object or the coordinates of a feature point of the object.

  A stereo measurement method is known as a method for measuring the position or shape of an object. The stereo measurement method uses an imaging device to image an object from a plurality of directions with different gaze directions, finds a corresponding point (measurement point) from the captured image, and obtains the position coordinates of the measurement point based on the principle of triangulation . In the stereo measurement method, it is necessary to detect where a measurement point image obtained by imaging a measurement point is located in a plurality of images.

  As a method for facilitating the detection of the measurement point image, the picked-up image picked up from two directions is divided by the frequency component of the video signal to obtain a multi-level image, and the image is further subjected to second-order differential processing to obtain a ternary value. Measurement points (feature points) are detected by performing the digitization process. And the method of comparing these picked-up images and detecting the place where a measurement point image respond | corresponds is proposed (for example, refer patent document 1).

Japanese Patent No. 3539788

  In various production sites, industrial robots are frequently used for work automation and labor saving. Such an industrial robot includes an imaging device such as a camera, images a work target (object), and performs image processing on the captured image to measure the position of the work target. The stereo measurement method described above may be used as a method for measuring the position of the work object. When the work object and the imaging device are moving relatively, motion blur occurs in the captured image. When motion blur occurs, there is a problem that it becomes difficult to specify a measurement point from a captured image.

  SUMMARY An advantage of some aspects of the invention is to solve at least a part of the problems described above, and the invention can be implemented as the following forms or application examples.

  (Application Example 1) A stage including a marker that places a measurement object and emits a plurality of regular reflected light, a light irradiation unit that irradiates light to the marker, and the measurement object And a plurality of imaging devices that capture the marker as the same image from different line-of-sight directions, and extract a marker image of the marker from the captured images captured by the plurality of imaging devices, from the marker image A point spread function calculating unit that calculates a point spread function; an image generating unit that generates a converted image by converting the captured image using the calculated point spread function; and And a position calculation unit that calculates the position of the measurement object.

  According to this configuration, the stage on which the measurement object is placed is provided with a marker that emits reflected light by the light irradiation unit that emits light. The reflected light emitted from the marker has regularity, that is, a characteristic that a reflection pattern is determined by the position of the marker or the imaging device with respect to the light irradiation unit. Therefore, it is possible to know the position information of the marker by examining the reflection pattern of the marker.

  In many cases, the picked-up image of the measurement object picked up by the image pickup device becomes a blurred image including so-called motion blur. In this captured image, a marker image composed of a reflection pattern is captured. By confirming the change in the reflection pattern of the marker image, it is possible to acquire the constituent elements of the point diffusion function such as the direction of movement and the change in velocity that cause the blur. The point spread function is used as a parameter that represents how the image is blurred. As a result, a point spread function can be calculated, and a non-blurred original image as a converted image can be generated from the captured image using the calculated point spread function. That is, it is possible to generate original images that are not blurred from blurred images captured by a plurality of imaging devices. Then, by applying a stereo measurement method using these unblurred original images (transformed images), the position of the object to be measured (the coordinates of the feature points) can be measured.

  (Application example 2) The position measuring device according to the above, wherein the marker has a plurality of reflecting surfaces provided three-dimensionally.

  According to this configuration, the pattern of the reflected light reflected from the marker can be varied depending on the position of the marker with respect to the light irradiation unit.

  (Application Example 3) The position measurement device according to the above, wherein the light irradiation unit irradiates the marker with light at an interval shorter than an exposure time of the imaging device.

  According to this configuration, the reflected light can be emitted from the marker a plurality of times within the exposure time of the imaging device. Therefore, a marker image composed of a predetermined number of reflection patterns can be formed in the captured image.

  Application Example 4 A pattern storage unit that stores in advance a reflection pattern including positional information of the marker is provided, and the point diffusion function calculation unit stores the captured marker image of the marker in the pattern storage unit. The position measuring apparatus described above, further comprising an image comparing unit that compares the reflected pattern.

  According to this configuration, it is possible to know the position of the marker when the image is captured by comparing the reflection pattern of the captured marker image with the reflection pattern having the marker position information stored in the pattern storage unit. it can.

  (Application Example 5) A measurement object placed on a stage having a marker that receives light irradiated from a light irradiation unit and reflects reflected light having regularity is different from the marker together with the marker by a plurality of imaging devices. An imaging step of imaging from the line-of-sight direction, a point diffusion function calculating step of extracting a marker image of the marker from the captured image captured in the imaging step, and calculating a point diffusion function from a pattern change of the marker image; An image generation step of generating a converted image by converting the captured image using the calculated point diffusion function, and a position calculation step of calculating the position of the object to be measured from the generated converted image. A position measurement method comprising:

  According to this method, the stage on which the measurement object is placed is provided with the marker that emits the reflected light by the light irradiation unit that irradiates the light. The reflected light emitted from the marker has regularity, that is, a characteristic that a reflection pattern is determined by the position of the marker or the imaging device with respect to the light irradiation unit. Therefore, it is possible to know the position information of the marker by examining the reflection pattern of the marker.

  In the imaging process, the captured image of the measured object is often a blurred image including so-called motion blur. The captured image includes a marker image made of a reflection pattern. By confirming the change of the reflection pattern included in the marker image, the marker image can acquire the constituent elements of the point diffusion function such as the direction of the motion causing the blur and the speed change. As a result, the point spread function calculation step can calculate the point spread function, and the image generation step uses the calculated point spread function to generate a non-blurred original image as a converted image from the captured image. Can do. That is, it is possible to generate original images that are not blurred from blurred images captured by a plurality of imaging devices. In the position calculation step, the position of the object to be measured (the coordinates of the feature points) can be measured by applying the stereo measurement method using the original image (converted image) that is not blurred.

  (Application example 6) A robot system characterized in that the position of a work object is measured by the position measuring device.

  According to this configuration, it is possible to provide a robot system that can easily calculate the position of the work object.

The figure which shows the structural example of the robot system to which a position measuring device is applied. The figure explaining the stage in which the workpiece | work which is a work target object is mounted. The figure which shows the relationship between an imaging device and the workpiece | work mounted on the stage. The flowchart which shows the flow of the position detection method of a workpiece | work. The figure which shows the picked-up image of the workpiece | work and marker imaged with the camera. The figure which shows the reproduced | regenerated original image as a conversion image. The figure explaining the stage of 2nd Example. The figure which shows the image of the workpiece | work concerning 2nd Example. The figure which shows the structure of the robot system to which the position measuring apparatus concerning a 1st modification is applied. The figure explaining the marker of the 2nd modification.

(First embodiment)
(About the configuration of the position measuring device)
An example in which the position measuring apparatus of this embodiment is applied to a robot system will be described with reference to FIGS. FIG. 1 is a diagram illustrating a configuration example of a robot system. FIG. 2 is a diagram illustrating a relationship between two imaging devices and a workpiece placed on a stage. FIG. 3 is a diagram for explaining a stage and a marker on which a workpiece as a workpiece and a work object are placed, (a) is a plan view of the stage, and (b) is an enlarged view of the marker. (C) is sectional drawing which shows the cross section of a marker.

  As shown in FIG. 1, a robot system 5 having a position measuring device 1 includes a robot 10, a workpiece placement unit 70 on which a workpiece W is placed, a first camera 20 and a second camera 21 as imaging devices. And a control unit 50. The robot 10 is a so-called articulated robot, and includes a base 25, arms 17A and 17B, a wrist portion 18 including a robot hand 19, and a first shaft 12 and a first shaft 12 which are joint portions that rotatably support each. It consists of two shafts 14 and a third shaft 16.

  The robot 10 is installed on the installation surface by the base 25. The base 25 includes the first rotation shaft 11 and can rotate the main body of the robot 10 above the base 25 about a vertical axis with respect to the installation surface. An arm 17 </ b> A is rotatably attached to the base 25 via the first shaft 12. The arm 17A includes a second rotating shaft 13, and the arm 17A itself can rotate in the axial direction. An arm 17B is rotatably attached to the arm 17A via the second shaft 14.

  The arm 17B includes a third rotating shaft 15, and the arm 17B itself can rotate in the axial direction. A wrist 18 is rotatably attached to the arm 17B via a third shaft 16. A robot hand 19 is provided at the tip of the wrist 18 and performs work on the workpiece W. The first to third shafts 12, 14, 16, the first to third rotating shafts 11, 13, 15 and the robot hand 19 are driven by a plurality of actuators (not shown) operated by a motor (not shown) or pneumatic equipment. , Configured to rotate. The plurality of actuators are driven based on a control signal sent from the control unit 50 via the cable 85.

  As shown in FIG. 2, the first camera 20 and the second camera 21 are juxtaposed on the wrist 18 of the robot 10 with a predetermined distance x. The first camera 20 and the second camera 21 function as two eyes of the robot 10, so that, for example, the direction in which the robot hand 19 is working is imaged with a predetermined exposure time by an imaging device such as a CCD. To do. Each image including the imaged workpiece W is sent to the control unit 50 via the cables 86 and 87 as an image signal.

  The work placing unit 70 includes a stage 72 on which the work W is placed. In stage 72, the robot 10 performs an operation on the workpiece W. The stage 72 has a stage surface 72a on which the workpiece W is placed, and the stage surface 72a includes one or more markers 90 formed of a plurality of reflecting surfaces having a function of reflecting light. Details of the marker 90 will be described later. For example, the work W is transported by a belt conveyor (not shown), is carried into the stage 72 of the work placing unit 70, and after the work by the robot 10 is performed, it is carried out to the next work placing unit 70 or the next process. . As shown in FIG. 2, the workpiece W is formed in, for example, a rectangular plate shape, and three through holes are provided in the plane portion. The workpiece placement unit 70 itself may be conveyed by a belt conveyor.

  The control unit 50 shown in FIG. 1 includes at least an image processing device 30, a pattern storage unit 33, a robot operation control unit 40, a light irradiation control unit 45, and a computer 60. The image processing device 30 includes at least a first image input unit 31, a second image input unit 32, and a position information acquisition unit 34, and is imaged by the first camera 20 and the second camera 21 of the robot 10. Then, the two images including the workpiece W and the marker 90 input via the first image input unit 31 and the second image input unit 32 are processed to acquire the position information of the workpiece W.

  The position information acquisition unit 34 includes an image comparison unit 35, a point spread function calculation unit 36, an image generation unit 37, and a position calculation unit 39, and performs various arithmetic processes to acquire the position information of the workpiece W. The operation of the position information acquisition unit 34 will be described later. The robot operation control unit 40 controls the operation of the robot 10 based on a command from the computer 60 and performs a predetermined operation on the workpiece W.

  The light irradiation control unit 45 includes a light irradiation unit 48 that irradiates light at a predetermined cycle to the marker 90 provided on the stage surface 72a on which the workpiece W is placed. The light irradiation part 48 is connected to the light irradiation control part 45 by the cable 88, and the period etc. which irradiate light are controlled. The pattern storage unit 33 moves the position of the marker 90 relative to the light irradiation unit 48 and the position of the imaging device (the first camera 20 or the second camera 21) relative to each other at a predetermined rate, while taking the imaging device. A plurality of marker images obtained by imaging the marker 90 using the are stored as a plurality of pattern data.

  The computer 60 functions as a central processing unit, and organically cooperates hardware resources such as a CPU, RAM, ROM, HDD, sequencer, and robot controller (not shown) with various software stored in the ROM, HDD, and the like. By operating, the robot 10, the image processing device 30, the light irradiation control unit 45, etc. are comprehensively controlled.

  Here, the stage 72 of this embodiment and the marker 90 provided on the stage 72 will be described with reference to FIG. As shown in FIG. 3A, the stage 72 is formed in a quadrangular shape. For example, a marker 90 is provided near one corner of the stage surface 72a. The marker 90 has a plurality of reflecting surfaces 92. The marker 90 preferably has a larger number of reflecting surfaces 92, but in this embodiment, for the sake of simplicity of description, for example, the marker 90 having three reflecting surfaces 92 will be described as an example.

  As shown in FIGS. 3B and 3C, the stage surface 72a is formed with, for example, an inverted triangular pyramid-shaped concave portion 91 having a bottom surface as an opening, and the concave surface 91 has three surfaces other than the opening. Triangular reflecting surfaces 92 having a function of reflecting light are respectively attached. By being formed in this way, when the marker 90 is irradiated with light from a predetermined direction, the marker 90 reflects light having a different reflection pattern depending on the light incident direction. In addition, it is preferable that the workpiece | work W is mounted avoiding the marker 90 provided in the stage surface 72a.

(About work in robot systems)
The operation using the above-described robot system 5 will be described with reference to FIGS. The robot 10 of the robot system 5 performs work on the workpiece W transferred to the stage 72 by a belt conveyor (not shown). In the present embodiment, no work W alignment device or positioning device is provided in order to improve production efficiency, generalize and simplify the equipment. For this reason, the workpieces W are transported in a random posture, and work unique to each is performed. Therefore, the robot 10 needs to detect the position and posture of the workpiece W that performs the work. That is, the robot system 5 detects the position of the workpiece W by imaging the workpiece W placed on the stage 72 using the first camera 20 and the second camera 21 shown in FIG.

  For example, it is assumed that the work W is placed on the stage 72 in a layout as shown in FIG. In order to detect the position of the workpiece W, the workpiece W on the stage 72 is imaged using the first camera 20 and the second camera 21 provided at the tip of the wrist portion 18 of the robot 10. At this time, it is preferable that the movement of the wrist portion 18 of the robot 10 is stopped or the workpiece W is stopped. However, in such a method, work and moving operations are interrupted, and work efficiency is significantly reduced. Therefore, in general, imaging is performed in a state where both are moving. At this time, the first camera 20 and the second camera 21 take images by opening the shutter over a predetermined exposure time.

  As a result, each captured image of the imaged workpiece W often becomes a so-called blurred image including motion blur. As it is, the position of the workpiece W cannot be detected. Therefore, some image processing is necessary. As an image processing method, for example, there is a method of restoring a non-blurred original image as a converted image by estimating how the blurred image is blurred and calculating an inverse characteristic of the blurred image.

(About workpiece position detection method)
Here, a workpiece position detection method using a point diffusion function as a parameter representing how the captured image is blurred will be described with reference to FIGS. FIG. 4 is a flowchart illustrating a flow of a workpiece position detection method, and FIG. 5 is a diagram illustrating captured images of a workpiece and a marker captured by a camera. FIG. 6 is a diagram illustrating a reproduced original image as a converted image.

  As shown in FIG. 4, this position detection method includes a workpiece carry-in step S11, a first marker & workpiece imaging step S12a, a second marker & workpiece imaging step S12b, a first image input step S13a, and a second image. Input step S13b, first image comparison step S14a, second image comparison step S14b, first point spread function calculation step S15a, second point spread function calculation step S15b, first image generation step S16a, A two-image generation step S16b and a workpiece position calculation step S17.

  In the workpiece carrying-in process S11, the workpiece W is conveyed by a belt conveyor or the like and supplied to the stage 72 of the workpiece placing unit 70 in the robot system 5 shown in FIG. At this time, since the workpiece W is supplied in a random posture, it is not known what posture the workpiece W is currently in. Therefore, the process proceeds to the next step.

  In the first marker and workpiece imaging step S12a, the workpiece W placed on the stage 72 is imaged using the first camera 20 shown in FIG. At this time, imaging is performed so that the marker 90 provided on the stage surface 72a on which the workpiece W is placed also falls within the same visual field. However, since the first camera 20 and the work W (stage 72) are relatively moving, the picked-up image to be picked up is so-called motion blur as shown in FIG. It becomes a blurred image. At this time, the light irradiation unit 48 shown in FIG. 1 irradiates the marker 90 with light at a cycle shorter than the exposure time of the first camera 20 by a control signal from the light irradiation control unit 45. Since the marker 90 has a plurality of reflection surfaces 92, it receives light from the light irradiation unit 48 and reflects the light. The reflected light forms a predetermined reflection pattern by the plurality of reflection surfaces 92.

  As a result, a blurred image B1 shown in FIG. The blurred image B1 includes an image Bw of the workpiece W including motion blur, and several reflection pattern images 105 of the marker 90 as the marker image 100 of the marker 90. In the example of FIG. 5, since light is irradiated four times from the light irradiation unit 48 during the exposure time of the first camera 20, four reflection pattern images 105a, 105b, 105c, and 105d are captured. In this embodiment, the distance Lab between the reflection pattern image 105a and the reflection pattern image 105b shown in FIG. 5B, the distance Lbc between the reflection pattern image 105b and the reflection pattern image 105c, and the distance Lcd between the reflection pattern image 105c and the reflection pattern image 105d. Are different.

In the second marker & workpiece imaging step S12b, the second camera 21 shown in FIG. 1 is used to image the workpiece W placed on the stage 72 in the same manner as in the first marker & workpiece imaging step S12a. As a result, a blur image B2 (not shown) is obtained. As shown in FIG. 2, the first camera 20 and the second camera 21 are arranged side by side at a distance x. Therefore, the obtained blurred image B2 is an image captured from a different viewpoint, in other words, from a different angle as compared to the blurred image B1.
In the following description, a case where the work W placed on the stage 72 is imaged using the first camera 20 will be described in order to simplify the description.

  In the first image input step S13a shown in FIG. 4, the blur image B1 including the marker image 100 captured by the first camera 20 is used as an image signal for the first of the image processing device 30 of the control unit 50 shown in FIG. The image is input into the image input unit 31. The captured blur image B1 is sent to the position information acquisition unit 34.

  In the first image comparison step S14a shown in FIG. 4, a group of a series of reflection patterns P stored in the pattern storage unit 33 shown in FIG. 1 and a marker of the marker 90 of the blurred image B1 captured by the position information acquisition unit 34 are used. The reflection pattern image 105 that is the image 100 is compared, and the reflection pattern P that is most approximate to the reflection pattern image 105 is extracted. As shown in FIG. 5C, for example, the reflection pattern Pa is extracted from the reflection pattern image 105a. It should be noted that the presence / absence or density of the meshes applied to the respective reflection surfaces 92 of the reflection pattern image 105a and the reflection pattern Pa in FIG. To express.

  As described above, the reflection pattern P is obtained by moving the marker 90 and the position of the first camera 20 relative to the light irradiation unit 48 shown in FIG. Therefore, the positional relationship between the marker 90 and the first camera 20 can be determined by confirming the light reflection state of the reflection surface 92 of the marker 90 in the reflection pattern P. Further, by making the position (coordinate system) of the first camera 20 known, the position (coordinate system) of the marker 90 can be known. In other words, the reflection pattern P has the position (coordinate) information of the marker 90. Therefore, by extracting the reflection pattern Pa close to the reflection pattern image 105a, the position (coordinates) of the marker 90 when the reflection pattern image 105a is imaged can be determined. This operation is performed for each of the reflection pattern images 105b, 105c, and 105d. As a result, the positions of the markers 90 when the reflection pattern images 105a, 105b, 105c, and 105d are imaged can be obtained.

  In the first point diffusion function calculation step S15a shown in FIG. 4, the marker image 100 (reflection pattern image 105) is extracted from the blurred image B1 to calculate the point diffusion function. The method for calculating the point spread function is not particularly limited. A known calculation method may be applied. Note that the marker image 100 has information on the spatial movement of the point spread function, that is, the direction and speed of the movement that causes the blur. In this case, it is possible to know the change in the position of the marker 90, that is, the motion vector (the motion direction in which the blur is generated) from the difference in the reflection pattern image 105. The marker 90 reflects light from the light irradiation unit 48 at substantially equal intervals (periods). Therefore, it is possible to know the relationship between the relative movement speed v between the first camera 20 and the marker 90 when the marker 90 is imaged from the distances Lab, Lbc, and Lcd between the reflection pattern images 105a, 105b, 105c, and 105d. it can.

  Taking the marker image 100 shown in FIG. 5B as an example, it can be seen that the marker 90 moves along the arrow M due to the change in the position (coordinates) of the marker 90 that can be seen from the difference in the reflection pattern image 105. The reflection pattern image 105 has a relationship of a distance Lbc between the reflection pattern image 105b and the reflection pattern image 105c> a distance Lcd between the reflection pattern image 105c and the reflection pattern image 105d> a distance Lab between the reflection pattern image 105a and the reflection pattern image 105b. ing. Therefore, the relative movement speed vab between the first camera 20 and the marker 90 when the reflection pattern image 105a and the reflection pattern image 105b are imaged, and the first when the reflection pattern image 105b and the reflection pattern image 105c are imaged. The relative movement speed vbc between the camera 20 and the marker 90 and the relative movement speed vcd between the first camera 20 and the marker 90 when imaging between the reflection pattern image 105c and the reflection pattern image 105d is relative movement. It can be seen that speed vbc> relative movement speed vcd> relative movement speed vab.

  Next, in a first image generation step S16a shown in FIG. 4, an original image (converted image) that is not blurred is generated from the blurred image B1 using the point diffusion function calculated in the first point diffusion function calculation step S15a. This method will be briefly described.

  The blur of the image is modeled as the following formula (1). As shown in this equation (1), when the original image is Image, the blurred image is Blur, and the point diffusion function that is a parameter indicating how the image is blurred is PSF, the blurred image (Blur) is point spread. It is given by the convolution operation of the function (PSF) and the original image (Image). As the name suggests, the point spread function (PSF) is a function that expresses how data that is originally one point is spread.

  If the Fourier transform operation for the function f is expressed as F [f] as in the following formulas (2), (3), and (4), the Fourier transform of the original image (Image) is an image, a blurred image. The Fourier transform of (Blur) can be expressed as Blur, and the Fourier transform of the point spread function (PSF) can be expressed as psf. Since the convolution operation is converted into simple multiplication by Fourier transform, Expression (1) is converted to Expression (5) below.

  As shown in the following equation (6), the Fourier transform (image) of the original image is the inverse of (psf) which is the Fourier transform of the point spread function (PSF) and the Fourier transform of the blurred image (Blur) (blur). ) And the product. When the inverse Fourier transform operation of the function f is IF [f], the original image (Image) without blur is the Fourier transform (image) of the original image obtained by Equation (6) as in Equation (7). It is obtained by inverse Fourier transform.

  The arithmetic processing as described above is a simple mathematical process, and an example thereof is a Wiener filter process. In this way, the original image G1 on which the unblurred work image Gw1 shown in FIG. 6A is displayed is generated.

  Also when the workpiece W placed on the stage 72 is imaged using the second camera 21, the second marker & workpiece imaging step S12b, the second image input step S13b, the second, as described above. The image comparison step S14b, the second point diffusion function calculation step S15b, and the second image generation step S16b are performed to generate the original image G2 on which the unblurred work image Gw2 shown in FIG. 6B is displayed. . The obtained original image G2 is an image viewed from a different viewpoint, in other words, from a different angle compared to the original image G1.

  Next, in a workpiece position calculation step S17 shown in FIG. 4, the position of the workpiece W is calculated based on the generated original image G1 and original image G2. A known position calculation method is applied as the position calculation method of the workpiece W. For example, the nature of the epipolar line, that is, the straight line connecting the first camera 20 and the measurement point is projected as a straight line to the second camera 21, and the point reflected in the first camera 20 is the second camera. In 21, the position of the workpiece W may be calculated by obtaining the coordinates of a plurality of measurement points using the property of being on the straight line (epipolar line). At this time, the measurement point is preferably determined by paying attention to the feature point of the shape of the workpiece W.

  And the information regarding the calculated position is sent to the robot operation control unit 40 shown in FIG. The robot operation control unit 40 controls the operation of the operation unit such as the arm 17 based on the information. That is, the robot operation control unit 40 calculates the drive amounts of a plurality of actuators for moving the robot 10 based on the information, generates a drive signal for each actuator, and sends the drive signal to each actuator. As a result, the arm 17 moves to a predetermined position, and a predetermined work is performed on the workpiece W by the robot hand 19.

The effects of this example will be described below.
(1) The robot system 5 to which the above-described position measuring device 1 is applied reflects reflected light on the stage 72 on which the workpiece W is placed at a cycle shorter than the exposure time of the first camera 20 and the second camera 21. A marker 90 composed of a plurality of reflecting surfaces 92 is provided. When measuring the position of the workpiece W (coordinates of feature points), the first camera 20 and the second camera 21 provided in the robot 10 are used to view the workpiece W and the marker 90 from the same line of sight in the same visual field. Take an image so that Since the stage 72 and the first camera 20 and the second camera 21 are relatively moved, the captured image of the workpiece W is a blurred image B including motion blur. The blurred image B includes a marker image 100 composed of the reflection pattern image 105 of the marker 90.

  Since the marker image 100 includes a plurality of reflection pattern images 105a, 105b, 105c, and 105d, the direction of the columns formed by the reflection pattern images 105a, 105b, 105c, and 105d and the reflection pattern images 105a, 105b, 105c, Constituent elements of the point spread function such as the direction of movement and the speed change that generate the blur can be easily obtained from the distances Lab, Lbc, and Lcd between 105d. As a result, the point spread function can be obtained easily and accurately simply by analyzing the captured images of the first camera 20 and the second camera 21, respectively, and image conversion is performed using the obtained point spread function. Thus, the original image G that is not blurred as the converted image can be accurately generated from the blurred image B as the captured image. And the position of a to-be-measured object can be measured more correctly by applying the stereo measurement method using these original images G. Therefore, the working efficiency of the robot system 5 can be improved.

(Second embodiment)
Here, an example in which the position measuring apparatus 1 of the second embodiment is applied to a robot system will be described with reference to FIGS. FIG. 7 is a diagram for explaining the stage of the second embodiment, and FIG. 8 is a diagram showing an image of the workpiece according to the second embodiment. In addition, 2nd Example is an example from which the installation condition of the marker 90 differs from 1st Example. Further, the same configurations and contents as those of the first embodiment are denoted by the same reference numerals and description thereof is omitted.

  As shown in FIG. 7, the stage 72 includes one marker 90a, 90b, and 90c in the vicinity of three corners of the stage surface 72a on which the workpiece W is placed. The markers 90a, 90b, 90c are formed in the same structure as the marker 90 described in the first embodiment.

  In the first marker & workpiece imaging step S12a and the second marker & workpiece imaging step S12b of the position detection method shown in FIG. 4, the first camera 20 and the second camera 21 shown in FIG. Imaging is performed so that the workpiece W and the markers 90a, 90b, and 90c to be placed are within the same visual field. As a result, a so-called blurred image Ba including motion blur shown in FIG. 8 is obtained. The blurred image Ba includes an image Bw of the workpiece W including motion blur, a marker image 110 of the marker 90a, a marker image 120 of the marker 90b, and a marker image 130 of the marker 90c near each of the three corners of the image. .

  The marker image 110 is formed, for example, from the lower right in FIG. 8 toward the upper left (arrow N1 direction). Therefore, it turns out that it is moving along arrow N1. Further, by paying attention to the bright spots 110a, 110b, 110c, and 110d of the blurred image Ba, the movement direction, the motion vector, and the relative movement between the first camera 20 and the marker 90a are obtained as in the first embodiment. It is possible to know the relationship of the speed v.

  The marker image 120 is formed from the lower right to the upper left (arrow N2 direction) in FIG. Therefore, it turns out that it is moving along arrow N2. The marker image 130 is formed from the lower right to the upper left (arrow N3 direction) in FIG. Therefore, it can be seen that the robot moves at least along the arrow N3. Further, when attention is paid to the distance L110 of the bright spots 110a to 110d of the marker image 110, the distance L120 of the bright spots 120a to 120d of the marker image 120, and the distance L130 of the bright spots 130a to 130d of the marker image 130, the distance L110 <distance L120 <. It can be seen that the distance between the bright points is similar and large because of the relationship of the distance L130. By analyzing changes in the direction of movement (arrows N1, N2, N3) and distances L110, L120, and L130, the robot rotates in the direction of arrow R in FIG. 8 while moving at the direction and speed determined from the marker image 110. You can see that

  That is, in the relative movement between the first camera 20 and the workpiece W (stage 72), even if motion blur occurs due to movement such as rotation or tilt, the point spread function is analyzed by analyzing the marker images 110, 120, and 130. Can be calculated more accurately. Therefore, the position of the workpiece W can be known more accurately from the obtained point spread function, and the work accuracy of the robot 10 can be improved. The number of markers 90 is not limited to one described in the first embodiment and three described in the second embodiment. What is necessary is just to determine the number of the markers 90 according to the motion of the 1st camera 20 and the 2nd camera 21, and the workpiece | work W (stage 72) anticipated.

  As mentioned above, although the Example of this invention was described, various deformation | transformation can be added with respect to the said Example in the range which does not deviate from the meaning of this invention. For example, modifications other than the above embodiment are as follows.

  (First Modification) In the above-described embodiment, the case where the first camera 20 and the second camera 21 are installed on the wrist 18 of the robot 10 has been described as an example, but the present invention is not limited to this. As shown in FIG. 9, which shows a configuration of a robot system to which the position measuring device 1 according to the first modification is applied, the first camera 20 and the second camera 21 are provided separately from the robot 10. It may be. Also in this case, it is possible to generate an original image (converted image) in which the motion blur accompanying the movement of the work W placed on the stage 72 is corrected.

  (Second Modification) In the above embodiment, the marker 90 having the three reflecting surfaces 92 has been described as an example, but the present invention is not limited to this. The marker 90 is preferably provided with more reflecting surfaces 92 in order to have more reflecting patterns. For example, as shown in FIG. 10, which is a diagram for explaining the marker of the second modification, brilliant-cut glass 94 used for diamond cutting may be embedded in the stage 72 to form the marker 90. In this case, the reflection pattern can be made finer, and the position information of the marker 90 can be reflected more finely in the reflection pattern. Therefore, it is possible to increase the calculation accuracy of the point spread function.

  (Third Modification) In the above-described embodiment, the case where the first camera 20 and the second camera 21 are applied as a plurality of imaging devices has been described as an example. However, the number of imaging devices is limited to two. It is not a thing and what is necessary is just two or more. Further, although the case where an articulated robot is employed as the robot 10 has been described, the present invention is not limited to this, and a scalar type robot may be used. Furthermore, the robot 10 can be used for various purposes such as gripping parts by the robot hand 19 and performing processing such as soldering and welding. Furthermore, the robot is not limited to an industrial robot, and may be a medical robot or a home robot.

  DESCRIPTION OF SYMBOLS 1 ... Position measuring device, 5 ... Robot system, 10 ... Robot, 18 ... Wrist part, 19 ... Robot hand, 20 ... 1st camera as an imaging device, 21 ... 2nd camera as an imaging device, 30 ... Image Processing device 31 ... first image input unit 32 ... second image input unit 33 ... pattern storage unit 34 ... position information acquisition unit 35 ... image comparison unit 36 ... point spread function calculation unit 37 ... image generation 39: Position calculation unit 45 ... Light irradiation control unit 48 ... Light irradiation unit 50 ... Control unit 60 ... Computer 72 ... Stage 90 90 Marker 92 92 Reflecting surface 100, 110, 120, 130 ... Marker images, 105a, 105b, 105c, 105d ... Reflection pattern images, B, B1, B2, Ba ... Blurred images as captured images, G, G1, G2 ... Original images as converted images, W ... Covered It works as a measurement object and the work object.

Claims (6)

A stage with a marker for placing a measurement object and emitting a plurality of reflected light with regularity;
A light irradiation unit for irradiating light to the marker;
A plurality of imaging devices for imaging the object to be measured and the marker as different images from different line-of-sight directions;
A point spread function calculator that extracts a marker image of the marker from the captured images captured by the plurality of imaging devices and calculates a point spread function from the marker image;
An image generation unit that generates a converted image by converting the captured image using the calculated point spread function;
A position calculation unit that calculates a position of the object to be measured from the generated converted image.   The position measuring device according to claim 1, wherein the marker is provided with a plurality of reflecting surfaces in a three-dimensional manner.   The position measuring device according to claim 1, wherein the light irradiation unit irradiates the marker with light at an interval shorter than an exposure time of the imaging device. A pattern storage unit that stores in advance a reflection pattern including position information of the marker,
4. The point spread function calculation unit includes an image comparison unit that compares the marker image of the captured marker with the reflection pattern stored in the pattern storage unit. The position measuring device according to claim 1. An object to be measured placed on a stage provided with a marker that receives light emitted from a light irradiation unit and reflects reflected light having regularity is imaged together with the marker from different gaze directions by a plurality of imaging devices. An imaging process to
A point diffusion function calculating step of extracting a marker image of the marker from the captured image captured in the imaging step and calculating a point diffusion function from a pattern change of the marker image;
An image generation step of generating a converted image by converting the captured image using the calculated point spread function;
A position calculation step of calculating a position of the measurement object from the generated converted image.   A robot system that measures the position of a work object with the position measurement device according to claim 1.

Images